[go: up one dir, main page]
Next Issue
Volume 13, July-1
Previous Issue
Volume 13, June-1
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
8.3
4.2
Journals
Remote Sensing
Volume 13
Issue 12
remotesensing-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Remote Sens., Volume 13, Issue 12 (June-2 2021) – 192 articles

Cover Story (view full-size image): Alaska’s Yukon River is a prominent feature of the Arctic–Boreal landscape—providing transportation and ecosystem services to many communities. We present a first analysis of satellite-derived snow properties and their interaction with hydrologic processes along the Yukon River, including the spring flood pulse and river ice break up (RIB) timing. A suite of passive microwave satellite-derived snow metrics, including Main Melt Onset Date, Snowoff Date, and Snowmelt Duration from 1988 to 2016, are presented and validated using in situ observations and complementary satellite data. We found meaningful correspondence between areal quantiles of the satellite snow metrics and measured streamflow quantiles and RIB observations, demonstrating the snow metrics’ potential for the monitoring and forecasting of hydrologic events along the Yukon River. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
22 pages, 4692 KiB  
Article
An Ontology-Based Approach to Enable Data-Driven Research in the Field of NDT in Civil Engineering
by Benjamí Moreno Torres, Christoph Völker, Sarah Mandy Nagel, Thomas Hanke and Sabine Kruschwitz
Remote Sens. 2021, 13(12), 2426; https://doi.org/10.3390/rs13122426 - 21 Jun 2021
Cited by 9 | Viewed by 4326
Abstract
Although measurement data from the civil engineering sector are an important basis for scientific analyses in the field of non-destructive testing (NDT), there is still no uniform representation of these data. An analysis of data sets across different test objects or test types [...] Read more.
Although measurement data from the civil engineering sector are an important basis for scientific analyses in the field of non-destructive testing (NDT), there is still no uniform representation of these data. An analysis of data sets across different test objects or test types is therefore associated with a high manual effort. Ontologies and the semantic web are technologies already used in numerous intelligent systems such as material cyberinfrastructures or research databases. This contribution demonstrates the application of these technologies to the case of the 1H nuclear magnetic resonance relaxometry, which is commonly used to characterize water content and porosity distribution in solids. The methodology implemented for this purpose was developed specifically to be applied to materials science (MS) tests. The aim of this paper is to analyze such a methodology from the perspective of data interoperability using ontologies. Three benefits are expected from this approach to the study of the implementation of interoperability in the NDT domain: First, expanding knowledge of how the intrinsic characteristics of the NDT domain determine the application of semantic technologies. Second, to determine which aspects of such an implementation can be improved and in what ways. Finally, the baselines of future research in the field of data integration for NDT are drawn. Full article
(This article belongs to the Special Issue Data Fusion, Integration and Advances of Non-destructive Testing Methods in Engineering and Geosciences)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The structure of a triple (Subject-property-Object), when related to a knowledge model (such as an ontology, <b>top</b>) and a set of data in tabular form (<b>bottom left</b>), provides a set of triples (<b>bottom center-right</b>).</p> Full article ">Figure 2
<p>Ontology classification sorted by level of abstraction of the respective fields. The examples in the middle column are ontologies grouped by ontology type. The right column presents examples of entities that could be part of an ontology of the corresponding level of abstraction.</p> Full article ">Figure 3
<p>Photograph of the NMR tomograph at BAM. Source: BAM.</p> Full article ">Figure 4
<p>NMR Principles, adapted version from [<a href="#B5-remotesensing-13-02426" class="html-bibr">5</a>]. (<b>a</b>)—Initial state in a water-containing medium without magnetization. (<b>b</b>)—Alignment of the protons when the sample is exposed to a static magnetic field. (<b>c</b>)—Deflection of the resulting magnetization into the transverse plane (x-y) by a short radiofrequency pulse. (<b>d</b>)—The magnetization relaxes back into the equilibrium state after termination of the radio-frequency pulse. (<b>e</b>)—Resulting measurable NMR signal (exponential decay of NMR amplitude). (<b>f</b>)—Relaxation time distribution as a result of a numerical inversion of the measured decaying signal. When converted using a material constant ρ, the x-axis can be converted to pore sizes.</p> Full article ">Figure 5
<p>Projected architecture development for the Mat-O-Lab initiative. Datasources refers to the source where raw data from experimental tests is stored. IDS stands for Integrated Data Storage: the different Rest-API allow raw-data to be transformed into triples (see <a href="#remotesensing-13-02426-f001" class="html-fig">Figure 1</a>) thanks to the connector; the triples are stored in the RDF-Triple Store.</p> Full article ">Figure 6
<p>Digital workflow. The domain expert (DE) is at the center of the process.</p> Full article ">Figure 7
<p>The goal of creating an endpoint is to automatize the populating process of the triplestore from the metadata files from experimental tests.</p> Full article ">Figure 8
<p>Digital workflow modified to allow ontology elicitation. Placing the ontology engineer (OE) at the center of the Test Description process only responds to the need to critically analyze and document the semantic transformation depicted in the <a href="#remotesensing-13-02426-f006" class="html-fig">Figure 6</a>.</p> Full article ">Figure 9
<p>Graphical representation of the knowledge base created from the <sup>1</sup>H NMR relaxometry test for humidity detection and porosity distribution. Some entities have been collapsed for better visualization. In the key the color codes are shown as encoded in the BWMD ontology file [<a href="#B105-remotesensing-13-02426" class="html-bibr">105</a>]. The shaded area 1 corresponds to the description of the content in the metadata file (zoom in <a href="#app1-remotesensing-13-02426" class="html-app">Appendix A</a> <a href="#remotesensing-13-02426-f0A1" class="html-fig">Figure A1</a>). The shaded area 2 corresponds to the description of the results data file (zoom in <a href="#app1-remotesensing-13-02426" class="html-app">Appendix A</a> <a href="#remotesensing-13-02426-f0A2" class="html-fig">Figure A2</a>). The shaded area 3 is a non-collapsed subset of area 1 that corresponds to the description of six variables of the NMR measurement used by the measurement machine software (zoom in <a href="#app1-remotesensing-13-02426" class="html-app">Appendix A</a> <a href="#remotesensing-13-02426-f0A3" class="html-fig">Figure A3</a>). It can be observed how each Quality is match to a xsd:Datatype (where the type of data to be stored is defined) and a rdfs:Literal (where the value of the Quality is stored).</p> Full article ">Figure A1
<p>Zoom on <a href="#remotesensing-13-02426-f009" class="html-fig">Figure 9</a> Area 1: Description of the content in the metadata file.</p> Full article ">Figure A2
<p>Zoom on <a href="#remotesensing-13-02426-f009" class="html-fig">Figure 9</a> Area 2: Description of the results data file.</p> Full article ">Figure A3
<p>Zoom on <a href="#remotesensing-13-02426-f009" class="html-fig">Figure 9</a> Area 3. description of six variables of the NMR measurement used by the measurement machine software.</p> Full article ">
20 pages, 18707 KiB  
Technical Note
Attention Multi-Scale Network for Automatic Layer Extraction of Ice Radar Topological Sequences
by Yiheng Cai, Dan Liu, Jin Xie, Jingxian Yang, Xiangbin Cui and Shinan Lang
Remote Sens. 2021, 13(12), 2425; https://doi.org/10.3390/rs13122425 - 21 Jun 2021
Cited by 1 | Viewed by 2293
Abstract
Analyzing the surface and bedrock locations in radar imagery enables the computation of ice sheet thickness, which is important for the study of ice sheets, their volume and how they may contribute to global climate change. However, the traditional handcrafted methods cannot quickly [...] Read more.
Analyzing the surface and bedrock locations in radar imagery enables the computation of ice sheet thickness, which is important for the study of ice sheets, their volume and how they may contribute to global climate change. However, the traditional handcrafted methods cannot quickly provide quantitative, objective and reliable extraction of information from radargrams. Most traditional handcrafted methods, designed to detect ice-surface and ice-bed layers from ice sheet radargrams, require complex human involvement and are difficult to apply to large datasets, while deep learning methods can obtain better results in a generalized way. In this study, an end-to-end multi-scale attention network (MsANet) is proposed to realize the estimation and reconstruction of layers in sequences of ice sheet radar tomographic images. First, we use an improved 3D convolutional network, C3D-M, whose first full connection layer is replaced by a convolution unit to better maintain the spatial relativity of ice layer features, as the backbone. Then, an adjustable multi-scale module uses different scale filters to learn scale information to enhance the feature extraction capabilities of the network. Finally, an attention module extended to 3D space removes a redundant bottleneck unit to better fuse and refine ice layer features. Radar sequential images collected by the Center of Remote Sensing of Ice Sheets in 2014 are used as training and testing data. Compared with state-of-the-art deep learning methods, the MsANet shows a 10% reduction (2.14 pixels) on the measurement of average mean absolute column-wise error for detecting the ice-surface and ice-bottom layers, runs faster and uses approximately 12 million fewer parameters. Full article
(This article belongs to the Special Issue The Cryosphere Observations Based on Using Remote Sensing Techniques)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The sites of data acquisition. The Arctic Archipelago is a group of islands along the Canadian Arctic Ocean; the glacial channels in the Arctic Archipelago are usually narrower than 3000 m.</p> Full article ">Figure 2
<p>Illustration of the imaging process of the MCoRDS. The topology slices are collected by three beams emitted by the MCoRDS. The flight direction of the aircraft is the along-track direction (range line), the cross-track direction (elevation angle bin) is perpendicular to the along-track direction and the fast-time axis (range bin) is perpendicular to both the along-rack direction and the cross-track direction.</p> Full article ">Figure 3
<p>Illustration of a topology slice with two kinds of labels, ice-surface layer and ice-bottom layer. White arrows indicate the ice-surface layer and ice-bottom layer. The red lines are the locations of ice-surface layer and ice-bottom layer (ground truth) marked by human expert.</p> Full article ">Figure 4
<p>Illustration of a segment of 3D radar topology sequences (<b>left</b>) and a 3D sub-glacial terrain (<b>right</b>) reconstructed from the sequence. H and W are the height and width of each topology slice, respectively, and D is the number of slices in a segment.</p> Full article ">Figure 5
<p>The whole framework of the MsANet. The MsANet, which is divided into 5 blocks, consists of improved C3D-M inserted with proposed Msks and 3D attention modules. The blue cuboid is the feature maps with a size of A × B × C (depth × width × height); for example, the first one after block 1 is the feature maps with a size of 5 × 32 × 32.</p> Full article ">Figure 6
<p>Improved C3D-M structure with 16, 32, 64 and 128 output channels per block. Numbers in brackets are the orders of convolution in blocks. Mp: mixed pooling.</p> Full article ">Figure 7
<p>Msk with K branches. The blue rectangle represents convolution and the green rectangle represents maximum pooling.</p> Full article ">Figure 8
<p>Illustration of the 3D attention module structure. The 3D attention module structure is made up of two branches, the upper branch is the 3D position attention module and the lower one is the 3D channel attention module. <math display="inline"><semantics> <mrow> <msup> <mi>E</mi> <mo>′</mo> </msup> <mtext> </mtext> <mi>and</mi> <mtext> </mtext> <mi>E</mi> </mrow> </semantics></math> are the feature maps with attention weights. In our attention module, the C8 unit, whose original placement position is indicated by the light-blue dashed box, used in [<a href="#B34-remotesensing-13-02425" class="html-bibr">34</a>] is deleted.</p> Full article ">Figure 9
<p>Visualization of radar topology slices. (<b>a</b>–<b>c</b>): Three set of comparison of extracted ice-surface layer positions and ice-bottom layer positions obtained by human labeled, method of Xu et al. [<a href="#B38-remotesensing-13-02425" class="html-bibr">38</a>], and method of MsANet. For better viewing, images are converted to gray-scale. The red line indicates the ice-surface layer and the green one is the ice-bottom layer. The width direction represents the elevation angle bins and the height direction represents the range bins, which is the same as the fast-time axis.</p> Full article ">Figure 10
<p>Visualization of comparisons of zoomed portions of the extraction results of Xu et al. [<a href="#B38-remotesensing-13-02425" class="html-bibr">38</a>] and the MsANet. (<b>a</b>–<b>c</b>): Three set of comparison of extracted zoomed ice-surface layer positions and ice-bottom layer positions obtained by human labeled, method of Xu et al. [<a href="#B38-remotesensing-13-02425" class="html-bibr">38</a>], and method of MsANet.</p> Full article ">Figure 11
<p>Reconstructions of terrain from results extracted by the MsANet without any interpolation method. Unit: meter. (<b>a</b>) Ice-surface layer. (<b>b</b>) Ice-bottom layer. The <span class="html-italic">x</span>-axis in each image means the flight path and the <span class="html-italic">y</span>-axis is the scanning width of the MCoRDS, while the color indicates the elevation of layers, which is the depth from the radar. For better observation, red boxes are used to mark the part where the MsANet extracted results are close to the ground truth and black boxes are applied to highlight the part where the MsANet extracted results are not similar to the ground truth.</p> Full article ">Figure 12
<p>Elevation differences of reconstructed layers using the results extracted by the MsANet. Unit: meter. (<b>a</b>) Elevation differences of the ice-surface layer. (<b>b</b>) Elevation differences of the ice-bottom layer. The topology slices between the two dashed lines in (<b>b</b>) are displayed on the right.</p> Full article ">
14 pages, 2356 KiB  
Communication
Integrating Ecological Assessments to Target Priority Restoration Areas: A Case Study in the Pearl River Delta Urban Agglomeration, China
by Xinchuang Chen, Feng Li, Xiaoqian Li, Hongxiao Liu, Yinhong Hu and Panpan Hu
Remote Sens. 2021, 13(12), 2424; https://doi.org/10.3390/rs13122424 - 21 Jun 2021
Cited by 11 | Viewed by 3003
Abstract
The identification and management of ecological restoration areas play important roles in promoting sustainable urban development. However, current research lacks a scientific basis for the scope and scale of ecological restoration. Further, the absence of a framework to assess policy goals and public [...] Read more.
The identification and management of ecological restoration areas play important roles in promoting sustainable urban development. However, current research lacks a scientific basis for the scope and scale of ecological restoration. Further, the absence of a framework to assess policy goals and public preferences that leads to identification of ecological restoration areas across the science-policy interface is difficult, and the existing frameworks’ performance has little applicability. We proposed a transdisciplinary framework to combine ecological quality, ecological health, and ecosystem services as an assessment endpoint to identify priority restoration areas. Further, we classified the ecological restoration areas on a township scale by K-means. Based upon policy goals and public preferences of the Pearl River Delta urban agglomeration, we chose air quality, biodiversity, soil fragility, recreation quality, ecosystem vigor, landscape metrics, and the water supply ecosystem service as elements of the evaluation system. This study showed that priority restoration areas accounted for 10.8% of the urban agglomeration area and classified township, largely in the difference between natural and semi-natural ecosystems and the human environment. Policymakers can use this framework comprehensively and flexibly to identify and classify ecological restoration areas to achieve policy goals and fulfil public preferences. Full article
(This article belongs to the Section Ecological Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location of the Pearl River Delta (PRD) urban agglomeration in China (<b>a</b>) and its land-use of PRD urban agglomeration in 2018 (<b>b</b>). DG: Dongguan, FS: Foshan, GZ: Guangzhou, HZ: Huizhou, JM: Jiangmen, SZ: Shenzhen, ZH: Zhuhai, ZQ: Zhaoqing, and ZS: Zhongshan.</p> Full article ">Figure 2
<p>A multi-disciplinary framework to identify ecological restoration areas by linking policy and ecological information. Identification of ecological restoration areas depends upon their ecological quality and health (black dotted bordered rectangle). The assessment of ecological quality and health should incorporate policy goals, local stakeholder needs, and expert knowledge (red dotted bordered rectangle) to select the desired index to ground the analysis in the given decision-making context to achieve equal and full representation. NPP: Net primary productivity.</p> Full article ">Figure 3
<p>Evaluation of hotspot indicators. We use a 10% threshold for ecological quality (EQ) and ecological health (EH). The total range of ecological quality and ecological health (shown in hyacinth) relative to hotspots (shown in blue).</p> Full article ">Figure 4
<p>Spatial results for the ecological hotspots and ecological restoration areas. (<b>a</b>) Ecological space quality hotspots; (<b>b</b>) ecological health hotspots; (<b>c</b>) ecological restoration areas, and (<b>d</b>) ecological restoration area in the city of the Pearl River Delta urban agglomeration.</p> Full article ">Figure 5
<p>The ecological management bundles based upon ecological restoration. (<b>a</b>) The spatial distribution characteristics of the ecological management bundle, (<b>b</b>) the ecological indices of ecological management bundles. Data were standardized to facilitate comparison among ecological indices. The length of each box is proportional to the relative abundance of the other ecological indices within each bundle. BD: biodiversity; SF: soil fragility; OR: outdoor recreation quality; V: vigor; O: organization; R: resilience, (<b>c</b>), The land-use types of ecological management bundles.</p> Full article ">
21 pages, 46739 KiB  
Article
A Pansharpening Generative Adversarial Network with Multilevel Structure Enhancement and a Multistream Fusion Architecture
by Liping Zhang, Weisheng Li, Hefeng Huang and Dajiang Lei
Remote Sens. 2021, 13(12), 2423; https://doi.org/10.3390/rs13122423 - 21 Jun 2021
Cited by 3 | Viewed by 2418
Abstract
Deep learning has been widely used in various computer vision tasks. As a result, researchers have begun to explore the application of deep learning for pansharpening and have achieved remarkable results. However, most current pansharpening methods focus only on the mapping relationship between [...] Read more.
Deep learning has been widely used in various computer vision tasks. As a result, researchers have begun to explore the application of deep learning for pansharpening and have achieved remarkable results. However, most current pansharpening methods focus only on the mapping relationship between images and the lack overall structure enhancement, and do not fully and completely research optimization goals and fusion rules. Therefore, for these problems, we propose a pansharpening generative adversarial network with multilevel structure enhancement and a multistream fusion architecture. This method first uses multilevel gradient operators to obtain the structural information of the high-resolution panchromatic image. Then, it combines the spectral features with multilevel gradient information and inputs them into two subnetworks of the generator for fusion training. We design a comprehensive optimization goal for the generator, which not only minimizes the gap between the fused image and the real image but also considers the adversarial loss between the generator and the discriminator and the multilevel structure loss between the fused image and the panchromatic image. It is worth mentioning that we comprehensively consider the spectral information and the multilevel structure as the input of the discriminator, which makes it easier for the discriminator to distinguish real and fake images. Experiments show that our proposed method is superior to state-of-the-art methods in both the subjective visual and objective assessments of fused images, especially in road and building areas. Full article
(This article belongs to the Section Remote Sensing Image Processing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Two kinds of finite difference operators. (<b>a</b>) first-level gradient operator, (<b>b</b>) second-level gradient operator.</p> Full article ">Figure 2
<p>Comparison of structural information. (<b>a</b>) PAN image, (<b>b</b>) horizontal structural information with first-level gradient operator, (<b>c</b>) vertical structural information with first-level gradient operator, (<b>d</b>) horizontal structural information with second-level gradient operator, (<b>e</b>) vertical structural information with second-level gradient operator.</p> Full article ">Figure 3
<p>Detailed architectures of the Generator.</p> Full article ">Figure 4
<p>Detailed architectures of the Discriminator.</p> Full article ">Figure 5
<p>Flow chart of reduced-resolution exmperiment.</p> Full article ">Figure 6
<p>Fusion results from the reduced-resolution experiment on the Gaofen-2 dataset.</p> Full article ">Figure 7
<p>Fusion results from the reduced-resolution experiment on the WorldView-2 dataset.</p> Full article ">Figure 8
<p>Residual images from the reduced-resolution experiment on the Gaofen-2 dataset.</p> Full article ">Figure 9
<p>Residual images from the reduced-resolution experiment on the WorldView-2 dataset.</p> Full article ">Figure 10
<p>Fusion results from the full-resolution experiment on the WorldView-2 dataset.</p> Full article ">Figure 11
<p>The performance of the land area in the reduced-resolution experiment.</p> Full article ">Figure 12
<p>The performance of the land area in the full-resolution experiment.</p> Full article ">Figure 13
<p>The performance of the vegetation area in the reduced-resolution experiment.</p> Full article ">Figure 14
<p>The performance of the vegetation area in the full-resolution experiment.</p> Full article ">Figure 15
<p>The performance of the building area in the reduced-resolution experiment.</p> Full article ">Figure 16
<p>The performance of the building area in the full-resolution experiment.</p> Full article ">Figure 17
<p>The performance of the road area in the reduced-resolution experiment.</p> Full article ">Figure 18
<p>The performance of the road area in the full-resolution experiment.</p> Full article ">
18 pages, 5608 KiB  
Article
A Case Study of the 3D Water Vapor Tomography Model Based on a Fast Voxel Traversal Algorithm for Ray Tracing
by Heng Hu, Min Liu, Jiqin Zhong, Xin Deng, Yunchang Cao and Peng Fang
Remote Sens. 2021, 13(12), 2422; https://doi.org/10.3390/rs13122422 - 21 Jun 2021
Cited by 5 | Viewed by 2374
Abstract
A fast voxel traversal algorithm for ray tracing was applied to build a 4 × 4 × 20 tomography model using the observation data of 11 ground-based Global Navigation Satellite System (GNSS) meteorology (GNSS/MET) stations in Hebei Province, China. The precipitation water vapor [...] Read more.
A fast voxel traversal algorithm for ray tracing was applied to build a 4 × 4 × 20 tomography model using the observation data of 11 ground-based Global Navigation Satellite System (GNSS) meteorology (GNSS/MET) stations in Hebei Province, China. The precipitation water vapor (PWV) observed at 05 a.m. (Universal Time Coordinated: UTC) on 10 December 2019, was used to reconstruct three-dimensional (3D) water vapor density fields over the test area. The tomographic results (GNSS_T) show that the water vapor density above this area is mainly below 25 g/m3 and is concentrated between the first to the fourth layers. The vertical distribution conforms to the exponential characteristics, while the horizontal distribution shows a decreasing trend from southwest to northeast. In addition, the results of the 0.25° grid dataset generated by the Global Forecast System (GFS) of the National Center for Environmental Forecasting (NCEP) (GFS_L) were interpolated to the height of the tomographic grid, which is in good agreement with the tomographic results. GFS_L is larger than GNSS_T on the first floor at the surface, with an average deviation of 0.19 g/m3. In contrast, GFS_L from the second floor to the top of the model is smaller than GNSS_T, with the average deviations distributed between −0.08 and −0.15 g/m3. Full article
(This article belongs to the Section Atmospheric Remote Sensing)
►▼ Show Figures

Figure 1

Figure 1
<p>Distribution of tomography stations.</p> Full article ">Figure 2
<p>Ray angle diagram (red line is the signal line).</p> Full article ">Figure 3
<p>The intersections of signal lines and the model (black and yellow balls are the positions of the station and satellite; blue, red, and green balls are the intersections of the signal lines and the model; red and blue lines are the signal and central lines).</p> Full article ">Figure 4
<p>Two dimensional demonstration of the fast voxel traversal algorithm for ray tracing. (The red dotted lines represent the signal extension line and the vertical line; R represents the distance between the starting point of the signal line and the origin of the X axis; T represents the distance between the starting point of the signal line and the first intersection of the x-axis direction; <math display="inline"><semantics> <mrow> <msub> <mi>X</mi> <mrow> <mi>v</mi> <mi>e</mi> <mi>c</mi> </mrow> </msub> </mrow> </semantics></math> represents the projection length of the signal line on the X axis).</p> Full article ">Figure 5
<p>Demonstration of the fast voxel traversal algorithm for ray tracing through the model grid (<b>a</b>) XY plane projection; (<b>b</b>) XZ plane projection; (<b>c</b>) a 3D display of rays passing through the model.</p> Full article ">Figure 6
<p>3D water vapor tomography result.</p> Full article ">Figure 7
<p>(<b>a</b>) GNSS_T of the first layer; (<b>b</b>) GFS_L of the first layer; (<b>c</b>) GNSS_T of the second layer; (<b>d</b>) GFS_L of the second layer; (<b>e</b>) GNSS_T of the third layer; (<b>f</b>) GFS_L of the third layer; (<b>g</b>) GNSS_T of the fourth layer; (<b>h</b>) GFS_L of the fourth layer.</p> Full article ">Figure 7 Cont.
<p>(<b>a</b>) GNSS_T of the first layer; (<b>b</b>) GFS_L of the first layer; (<b>c</b>) GNSS_T of the second layer; (<b>d</b>) GFS_L of the second layer; (<b>e</b>) GNSS_T of the third layer; (<b>f</b>) GFS_L of the third layer; (<b>g</b>) GNSS_T of the fourth layer; (<b>h</b>) GFS_L of the fourth layer.</p> Full article ">Figure 8
<p>(<b>a</b>) Comparison of GNSS_T and GFS_L between 20 layers at following stations: szcz, szhj, szbd, szzz, szme, szcf. (<b>b</b>) Comparison of GNSS_T and GFS_L between 20 layers at following stations: szag, szyx, szyq, szax, szwe.</p> Full article ">Figure 9
<p>The deviation between the water vapor density of the tomography and GFS data.</p> Full article ">Figure 10
<p>(<b>a</b>) GNSS_T of the ninth layer; (<b>b</b>) GFS_L of the ninth layer; (<b>c</b>) Deviation between GNSS_T and GFS_L of the ninth layer; (<b>d</b>) GNSS_T of the thirteenth layer; (<b>e</b>) GFS_L of the thirteenth layer; (<b>f</b>) Deviation between GNSS_T and GFS_L of the thirteenth layer.</p> Full article ">Figure 10 Cont.
<p>(<b>a</b>) GNSS_T of the ninth layer; (<b>b</b>) GFS_L of the ninth layer; (<b>c</b>) Deviation between GNSS_T and GFS_L of the ninth layer; (<b>d</b>) GNSS_T of the thirteenth layer; (<b>e</b>) GFS_L of the thirteenth layer; (<b>f</b>) Deviation between GNSS_T and GFS_L of the thirteenth layer.</p> Full article ">
21 pages, 8721 KiB  
Article
Impacts of Reservoir Water Level Fluctuation on Measuring Seasonal Seismic Travel Time Changes in the Binchuan Basin, Yunnan, China
by Chunyu Liu, Hongfeng Yang, Baoshan Wang and Jun Yang
Remote Sens. 2021, 13(12), 2421; https://doi.org/10.3390/rs13122421 - 21 Jun 2021
Cited by 9 | Viewed by 2624
Abstract
An airgun source in a water reservoir has been developed in the past decade as a green active source that had been proven effective to derive short-term subsurface structural changes. However, seasonal water level fluctuation in the reservoir affects the airgun signal, and [...] Read more.
An airgun source in a water reservoir has been developed in the past decade as a green active source that had been proven effective to derive short-term subsurface structural changes. However, seasonal water level fluctuation in the reservoir affects the airgun signal, and thus whether the airgun signals can be used to derive robust seasonal variation in subsurface structure remains unclear. We use the airgun data observed in the Binchuan basin to estimate the seasonal variation of seismic travel time and compare the results with those derived from ambient noise data in the same frequency band. Our main observation is that seasonal change δt/t from airgun is negatively correlated to the variation of dominant frequency and water table fluctuation in the reservoir. One possible explanation is that water table fluctuation in the reservoir affects the dominant frequency of the airgun signal and causes significant phase shift. We also compute the travel time changes in P-wave from the empirical Green’s function after deconvolving the waveforms from a reference station that is 50 m from the airgun source. The dominant frequency after deconvolution still shows seasonal variation and correlates inversely to the travel time changes, suggesting that deconvolution cannot completely eliminate the source effect on travel time changes. We also use ambient noise cross-correlation to retrieve coda waves and then derive travel time changes in monthly stacked cross-correlations relative to a yearly average cross-correlation. We observe that seismic travel time increases to its local maximum in the end of August. The travel time changes lag behind the precipitation for about one month. We apply a poroelastic physical model to explain seismic travel time changes and find that a combined effect from precipitation and evaporation might induce the seasonal changes as shown in the ambient noise data. However, the pattern of travel time changes from the airgun differs from that from ambient noise, reflecting the strong effects of airgun source property changes. Therefore, we should be cautious to derive long-term subsurface structural variation from the airgun source and put more attention on stabilizing the dominant frequency of each excitation in the future experiments. Full article
(This article belongs to the Special Issue Advances in Seismic Interferometry)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Short-period stations (red triangle) installed during the airgun source excitation experiment. Red text above the triangle denotes the station name. The black stars mark the location of the airgun array.</p> Full article ">Figure 2
<p>Original near-field waveform (<b>a</b>) recorded by a reference station CKT. The signal from 0 to 0.3 s is the primary pulse (gray area marked as primary). The signal from 0.6 to 2 s is bubble pulse. (<b>b</b>) The spectrogram of the waveform in (A). The primary pulse is in high energy and in a wide bandwidth, e.g., 7–30 Hz. The bubble pulse is in a relatively low frequency bandwidth, 2–6 Hz.</p> Full article ">Figure 3
<p>Seasonal variation of the dominant frequency (red) and P-wave signal amplitude (blue) for the reference station (<b>a</b>) and other receivers (<b>b</b>,<b>c</b>). They decrease in the summer and correlate to each other.</p> Full article ">Figure 4
<p>Airgun signals within 30 km from the excitation site. We observe P-wave arrivals and S-wave arrivals with apparent P- and S-wave velocities of 5.5 km/s and 2.8 km/s, respectively. Text near the tail of each signal denotes the station name.</p> Full article ">Figure 5
<p>Waveform cross section of vertical components at the reference station, CKT (<b>a</b>), and one receiver 53278 (<b>c</b>). Waveforms are aligned with the original time of each shot (<b>b</b>,<b>d</b>). Strong clock drifts in the reference station and receivers start around days 230.</p> Full article ">Figure 6
<p>Computation of the rectilinearity of one shot at the station 53278. (<b>a</b>) Three components of one airgun shot. (<b>b</b>) Correlation coefficient between a P-wave template and the vertical SHZ component. (<b>c</b>) Maximum value of differential of rectilinearity constrains the rough arrival time of P-wave. (<b>d</b>) Rectilinearity of the vertical component. The rectilinearity of P-wave is close to 1.</p> Full article ">Figure 7
<p>Procedures of the moving-window-cross-spectrum to compute the delay time. (<b>a</b>) Airgun segments center at 0.5 s with length of 0.6 s from the reference and a current airgun signal of station CKT in the passband of 2 to 6 Hz. The gray area marks the selected segments. (<b>b</b>) The cross-coherency between two selected signals. (<b>c</b>) Phase change over the interested frequency range. The slope determines the delay time, and the intercept constrains a constant phase shift. We can estimate a delay time <math display="inline"><semantics> <mrow> <mi>δ</mi> <mi>t</mi> </mrow> </semantics></math> from the frequency-dependent phase shift <math display="inline"><semantics> <mrow> <mi>ϕ</mi> <mo>/</mo> <mn>2</mn> <mi>π</mi> <mi>f</mi> </mrow> </semantics></math>. We can also compute a frequency-independent phase shift <math display="inline"><semantics> <mrow> <msub> <mi>ϕ</mi> <mi>c</mi> </msub> </mrow> </semantics></math> from the intercept of the linear regression. (<b>d</b>) The constant phase shift <math display="inline"><semantics> <mrow> <msub> <mi>ϕ</mi> <mi>c</mi> </msub> </mrow> </semantics></math> over different moving window. Steps in (<b>a</b>–<b>c</b>) estimate one point centering at 0.5 s. Looping through different moving window, we can estimate constant phase shift in different travel time <math display="inline"><semantics> <mrow> <msub> <mi>ϕ</mi> <mi>c</mi> </msub> <mo>/</mo> <mi>t</mi> </mrow> </semantics></math>. (<b>e</b>) Similar to (<b>d</b>) but for delay time.</p> Full article ">Figure 8
<p>Seismic travel time change for all stations. The <math display="inline"><semantics> <mrow> <mi>δ</mi> <mi>t</mi> <mo>/</mo> <mi>t</mi> </mrow> </semantics></math> with error bar increases from −5% in the winter to 5% in the summer. The error bars represent standard deviation of <math display="inline"><semantics> <mrow> <mi>δ</mi> <mi>t</mi> <mo>/</mo> <mi>t</mi> </mrow> </semantics></math> in one day.</p> Full article ">Figure 9
<p>P-wave seismic travel time change from empirical Green’s functions. After the deconvolution, we observe that the dominant frequency of P-wave in these empirical Green’s functions varies and correlate inversely to the delay time change.</p> Full article ">Figure 10
<p>Comparison of <math display="inline"><semantics> <mrow> <mi>δ</mi> <mi>t</mi> <mo>/</mo> <mi>t</mi> </mrow> </semantics></math> before and after the deconvolution.</p> Full article ">Figure 11
<p>dt/t(%) measurements (<b>b</b>) from ambient noise among CKT, 53265 and 53268 (<b>a</b>) and dt/t(%) measurements from airgun at the station 53268 (<b>c</b>).</p> Full article ">Figure 12
<p>Investigation of mechanism for seasonal changes in seismic travel time from ambient noise. Seismic travel time changes do not obviously correlate to the water table fluctuation of the reservoir (<b>a</b>), local precipitation (<b>c</b>), air pressure (<b>b</b>), and temperature (<b>d</b>). (<b>e</b>) Seismic travel time changes modeled from a poroelastic physical model match with the observed ones. (<b>f</b>) The optimal diffusivity is 0.01 m<sup>2</sup>/s for a minimum residual between synthetic and observed seismic velocity changes.</p> Full article ">Figure 13
<p><math display="inline"><semantics> <mrow> <mi>δ</mi> <mi>t</mi> <mo>/</mo> <mi>t</mi> </mrow> </semantics></math> at the station 53278 is inversely correlated to the dominant frequency of P-wave from airgun (<b>a</b>) and the water table fluctuation (<b>b</b>).</p> Full article ">
25 pages, 9869 KiB  
Article
DLR Earth Sensing Imaging Spectrometer (DESIS) Level 1 Product Evaluation Using RadCalNet Measurements
by Mahesh Shrestha, Dennis Helder and Jon Christopherson
Remote Sens. 2021, 13(12), 2420; https://doi.org/10.3390/rs13122420 - 21 Jun 2021
Cited by 8 | Viewed by 3688
Abstract
The DLR Earth Sensing Imaging Spectrometer (DESIS) is the first hyperspectral imaging spectrometer installed in the Multi-User System for Earth Sensing (MUSES) on the International Space Station (ISS) for acquiring routine science grade images from orbit. It was launched on 29 June 2018 [...] Read more.
The DLR Earth Sensing Imaging Spectrometer (DESIS) is the first hyperspectral imaging spectrometer installed in the Multi-User System for Earth Sensing (MUSES) on the International Space Station (ISS) for acquiring routine science grade images from orbit. It was launched on 29 June 2018 and integrated into MUSES. DESIS measures energy in the spectral range of 400 to 1000 nm with high spatial and spectral resolution: 30 m and 2.55 nm, respectively. DESIS data should be sufficiently quantitative and accurate to use it for different applications and research. This work performs a radiometric evaluation of DESIS Level 1 product (Top of Atmosphere (TOA) reflectance) by comparing it with coincident Radiometric Calibration Network (RadCalNet) measurements at Railroad Valley Playa (RVUS), Gobabeb (GONA), and La Crau (LCFR). RVUS, GONA, and LCFR offer 4, 15, and 5 coincident datasets between DESIS and RadCalNet measurements, respectively. The results show an agreement between DESIS and RadCalNet TOA reflectance within ~5% for most spectral regions. However, there is an additional ~5% disagreement across the wavelengths affected by water vapor absorption and atmospheric scattering. Among the three RadCalNet sites, RVUS and GONA show a similar measurement disagreement with DESIS of ~5%, while LCFR differs by ~10%. Agreement between DESIS and RadCalNet measurements is variable across all three sites, likely due to surface type differences. DESIS and RadCalNet agreement show a precision of ~2.5%, 4%, and 7% at RVUS, GONA, and LCFR, respectively. RVUS and GONA, which have a similar surface type, sand, have a similar level of radiometric accuracy and precision, whereas LCFR, which consists of sparse vegetation, has lower accuracy and precision. The observed precision of DESIS Level 1 products from all the sites, especially LCFR, can be improved with a better Bidirectional Reflection Distribution Function (BRDF) characterization of the RadCalNet sites. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>DESIS image of Railroad Valley Playa (<b>a</b>) and site view (<b>b</b>).</p> Full article ">Figure 2
<p>DESIS image of Gobabeb (<b>a</b>) and site view (<b>b</b>).</p> Full article ">Figure 3
<p>DESIS image of La Crau (<b>a</b>) and site view (<b>b</b>).</p> Full article ">Figure 4
<p>DESIS image of Baotau (<b>a</b>) and site view (<b>b</b>).</p> Full article ">Figure 5
<p>Acquisition geometry on RVUS. Blue symbols represent DESIS viewing geometry, red symbols represent DESIS solar geometry, and black symbols represent RadCalNet solar geometry.</p> Full article ">Figure 6
<p>RVUS TOA reflectance, RadCalNet (<b>a</b>) and DESIS (<b>b</b>). RadCalNet and DESIS individual spectrum comparison (<b>c</b>–<b>f</b>). DESIS acquisition dates and RadCalNet measurement dates are expressed in (yyyy_day of year).</p> Full article ">Figure 6 Cont.
<p>RVUS TOA reflectance, RadCalNet (<b>a</b>) and DESIS (<b>b</b>). RadCalNet and DESIS individual spectrum comparison (<b>c</b>–<b>f</b>). DESIS acquisition dates and RadCalNet measurement dates are expressed in (yyyy_day of year).</p> Full article ">Figure 7
<p>DESIS normalized TOA reflectance on RVUS (<b>a</b>). Absolute TOA difference between DESIS and RadCalNet TOA reflectance (<b>b</b>).</p> Full article ">Figure 8
<p>Mean and standard deviation of 4 normalized DESIS TOA reflectance datasets shown in <a href="#remotesensing-13-02420-f007" class="html-fig">Figure 7</a>a.</p> Full article ">Figure 9
<p>DESIS and RadCalNet acquisition geometry on GONA. The blue symbols represent DESIS viewing geometry, the red symbols represent DESIS solar geometry, and the black symbols represent RadCalNet solar geometry.</p> Full article ">Figure 10
<p>GONA TOA reflectance: RadCalNet (<b>a</b>) and DESIS (<b>b</b>).</p> Full article ">Figure 11
<p>DESIS normalized TOA reflectance on GONA (<b>a</b>). Absolute TOA difference between DESIS and RadCalNet TOA reflectance (<b>b</b>).</p> Full article ">Figure 12
<p>Mean and standard deviation of 15 normalized DESIS TOA reflectance shown in <a href="#remotesensing-13-02420-f011" class="html-fig">Figure 11</a>a.</p> Full article ">Figure 13
<p>DESIS and RadCalNet acquisition geometry on LCFR. The blue and red symbols represent DESIS viewing and solar geometry, respectively. The black symbols represent RadCalNet solar geometry.</p> Full article ">Figure 14
<p>LCFR TOA reflectance, RadCalNet (<b>a</b>) and DESIS (<b>b</b>).</p> Full article ">Figure 15
<p>DESIS normalized TOA reflectance on LCFR (<b>a</b>). Absolute TOA difference between DESIS and RadCalNet TOA reflectance (<b>b</b>).</p> Full article ">Figure 16
<p>Mean and standard deviation of 6 normalized DESIS TOA reflectance shown in <a href="#remotesensing-13-02420-f015" class="html-fig">Figure 15</a>a.</p> Full article ">Figure 17
<p>Mean and standard deviation of normalized DESIS TOA reflectance from all sites.</p> Full article ">Figure 18
<p>Mean of normalized DESIS TOA reflectance using different sites (<b>a</b>) and its corresponding uncertainty (<b>b</b>).</p> Full article ">Figure 19
<p>Mean of normalized DESIS TOA reflectance using different sites (<b>a</b>) and its corresponding uncertainty (<b>b</b>). The blue, green, and red data represent RVUS, GONA, and LCFR data, respectively.</p> Full article ">Figure 20
<p>Mean of normalized DESIS TOA reflectance on GONA. Blue and magenta curves represent spectra from descending and ascending mode acquisitions, respectively.</p> Full article ">
20 pages, 8302 KiB  
Article
A More Accurate Field-to-Field Method towards the Wind Retrieval of HY-2B Scatterometer
by Xinjie Shi, Boheng Duan and Kaijun Ren
Remote Sens. 2021, 13(12), 2419; https://doi.org/10.3390/rs13122419 - 21 Jun 2021
Cited by 5 | Viewed by 2206
Abstract
In this paper, we present a method for retrieving sea surface wind field (SSWF) from HaiYang-2B (HY-2B) scatterometer data. In contrast to the conventional algorithm, i.e., using a point-to-point (P2P) method based on geophysical model functions (GMF) to retrieve SSWF by spaceborne scatterometer, [...] Read more.
In this paper, we present a method for retrieving sea surface wind field (SSWF) from HaiYang-2B (HY-2B) scatterometer data. In contrast to the conventional algorithm, i.e., using a point-to-point (P2P) method based on geophysical model functions (GMF) to retrieve SSWF by spaceborne scatterometer, we introduce a more accurate field-to-field (F2F) retrieval method based on convolutional neural network (CNN). We fully consider the spatial correlation and continuity between adjacent observation points, and input the observation data of continuous wind field within a certain range into the neural network to construct the neural network model, and then synchronously obtain the wind field within the range. The wind field obtained by our retrieval method maintains its continuity and solves the problem of ambiguity removal in traditional wind direction retrieval methods. Comparing the retrieval results with the buoy data, the results show that the root mean square errors (RMSE) of wind direction and wind speed are less than 0.18 rad (10.31°) and 0.75 m/s, respectively. The retrieval accuracy is better than the L2B product of HY-2B. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Locations of TAO (blue) and NDBC (red) buoys used in this paper.</p> Full article ">Figure 2
<p>Schematic representation of two NN-based approaches: P2P (<b>a</b>) and F2F (<b>b</b>). What is happening in grey picture? The P2P retrieval method inputs the parameters of a single vector cell (NRCS, etc.) into the P2P TF to obtain the corresponding single wind field (wind speed and direction). The results obtained by the P2P method lack continuity and there are ambiguous solutions. The F2F method inputs the parameters of multiple wind vector cells (m × n in (<b>b</b>)) within a certain range into the F2F TF, and retrieve m × n wind fields at the same time. The F2F TF extracts the spatial continuity characteristics of the wind field and applies it to the retrieval process. At the same time, the entire continuous wind field composed of multiple wind vector cells is obtained, which can fundamentally eliminate the ambiguous solutions, and the wind vector cells of the obtained wind field are smoother and more continuous. The F2F TF is a retrieval model, which contains mathematical formulas for the process (retrieval) from the measured parameter data of the scatterometer to the wind field data. The retrieval process is completed by a neural network, and the continuity characteristics of the wind field will be applied to this process. TF represents the transfer function. m × n is the base size, i.e., the number of cells that serve as NN inputs and/or outputs. The case for m = n = 3 is shown.</p> Full article ">Figure 3
<p>Wind direction statistics (<b>a</b>) and wind speed statistics (<b>b</b>). In (<b>a</b>), the interval of the abscissa of the wind direction is 10°, starting from 0, every 10° is divided into one category for statistics. In (<b>b</b>), starting from 0, every 1 m/s is counted as a category.</p> Full article ">Figure 4
<p>The schematic diagram of our CNN. The structure and parameters of the wind speed neural network are different from the wind direction neural network. The structure and parameter settings of the two neural networks are shown in the figure (The data in the brackets are the structural parameters of the wind direction neural network).</p> Full article ">Figure 5
<p>Variations of the wind direction RMSE (<b>a</b>) and wind speed RMSE (<b>b</b>).</p> Full article ">Figure 6
<p>Comparison of label data and F2F-CNN method results (<b>a</b>), label data and HY-2B L2B results (<b>b</b>). It can be clearly seen that compared to the HY-2B L2B data, the F2F-CNN wind speed is more confirmed to the ECMWF ERA5 wind speed.</p> Full article ">Figure 7
<p>(<b>a</b>,<b>d</b>) show the wind speeds of ECMWF below 10 m/s and above 10 m/s, respectively. (<b>b</b>,<b>e</b>) show the difference map between F2F-CNN wind speed and ECMWF wind speed. (<b>c</b>,<b>f</b>) show the difference map between HY-2B wind speed and ECMWF wind.</p> Full article ">Figure 8
<p>Comparison of F2F-CNN method results and ECMWF wind direction (<b>a</b>), HY-2B L2B results and ECMWF wind direction (<b>b</b>). It can be clearly seen that results of the F2F-CNN method are better fitted to the ECMWF wind direction and poses to smaller deviations.</p> Full article ">Figure 9
<p>(<b>a</b>,<b>c</b>,<b>e</b>) show the comparison between the F2F-CNN wind direction and the ECMWF wind direction, and (<b>b</b>,<b>d</b>,<b>f</b>) show the comparison between the HY-2B wind direction and the ECMWF wind direction.</p> Full article ">Figure 10
<p>(<b>a</b>) shows the absolute difference between the F2F-CNN wind speed and the ECMWF data, and (<b>b</b>) shows the absolute difference between the HY-2B L2B wind speed and the ECMWF data. It is obvious seen from the figure that the wind speed of the HY-2B L2B data at the center of the typhoon is significantly higher than the ECMWF value, and the overall wind speed error of the F2F-CNN is more uniform and smaller.</p> Full article ">Figure 11
<p>The wind direction map derived by F2F-CNN vs. Target (<b>a</b>) and HY-2B L2B vs. Target (<b>b</b>).</p> Full article ">Figure 12
<p>The wind speed and wind direction from F2F-CNN (<b>a</b>), HY-2B L2B (<b>b</b>) and ECMWF (<b>c</b>).</p> Full article ">
22 pages, 23060 KiB  
Article
A Spatial-Scale Evaluation of Soil Consolidation Concerning Land Subsidence and Integrated Mechanism Analysis at Macro-, and Micro-Scale: A Case Study in Chongming East Shoal Reclamation Area, Shanghai, China
by Qingbo Yu, Xuexin Yan, Qing Wang, Tianliang Yang, Wenxi Lu, Meng Yao, Jiaqi Dong, Jiewei Zhan, Xinlei Huang, Cencen Niu and Kai Zhou
Remote Sens. 2021, 13(12), 2418; https://doi.org/10.3390/rs13122418 - 21 Jun 2021
Cited by 12 | Viewed by 3115
Abstract
Land reclamation has been increasingly employed in many coastal cities to resolve issues associated with land scarcity and natural hazards. Especially, land subsidence is a non-negligible environmental geological problem in reclamation areas, which is essentially caused by soil consolidation. However, spatial-scale evaluation on [...] Read more.
Land reclamation has been increasingly employed in many coastal cities to resolve issues associated with land scarcity and natural hazards. Especially, land subsidence is a non-negligible environmental geological problem in reclamation areas, which is essentially caused by soil consolidation. However, spatial-scale evaluation on the average degree of consolidation (ADC) of soil layers and the effects of soil consolidation on land subsidence have rarely been reported. This study aims to carry out the integrated analysis on soil consolidation and subsidence mechanism in Chongming East Shoal (CES) reclamation area, Shanghai, at spatial-, macro-, and micro-scale so that appropriate guides can be provided to resist the potential environmental hazards. The interferometric synthetic aperture radar (InSAR) technique was utilized to retrieve the settlement curves of the selected onshore (Ra) and offshore (Rb) areas. Then, the hyperbolic (HP) model and three-point modified exponential (TME) model were combined applied to predict the ultimate settlement and to determine the range of ADC rather than a single pattern. With two boreholes Ba and Bb set within Ra and Rb, conventional tests, MIP test, and SEM test were conducted on the collected undisturbed soil to clarify the geological features of exposed soil layers and the micro-scale pore and structure characteristics of representative compression layer. The preliminary results showed that the ADC in Rb (93.1–94.1%) was considerably higher than that in Ra (60.8–78.7%); the clay layer was distinguished as the representative compression layer; on micro-scale, the poor permeability conditions contributed to the low consolidation efficiency and slight subsidence in Rb, although there was more compression space. During urbanization, the offshore area may suffer from potential subsidence when it is subjected to an increasing ground load, which requires special attention. Full article
(This article belongs to the Special Issue Applications of Remote Sensing in Geological Engineering)
►▼ Show Figures

Figure 1

Figure 1
<p>A sketch of the expansion of Shanghai’s coastal land. (<b>a</b>) The widened Yangtze River; (<b>b</b>) the joint effects of seawater support and flocculation; (<b>c</b>) the expanding new lands; (<b>d</b>) the vessels on the voyage; (<b>e</b>) hydraulic reclamation; (<b>f</b>) multiple soil layers.</p> Full article ">Figure 2
<p>Google Earth-based overview of Chongming East Shoal (CES). (<b>a</b>) The main reclamation areas in Shanghai; (<b>b</b>) the advancement of multi-phase reclamation projects in CES since the 1990s; (<b>c</b>) site environment and land use; (<b>d</b>) SBAS-derived vertical deformation map from 22 March 2015 to 2 December 2019 (data from Reference [<a href="#B27-remotesensing-13-02418" class="html-bibr">27</a>]); (<b>e</b>) comparison of SBAS-derived deformation velocity (ν<sub>S</sub>) and fitting leveling data (ν<sub>L</sub>) at the leveling location (data from Reference [<a href="#B27-remotesensing-13-02418" class="html-bibr">27</a>]). (Note that the leveling measurement followed the specification of the second-order leveling with an error of 2 mm, based on the Chinese national height data from 1985).</p> Full article ">Figure 3
<p>Stratigraphic structure of boreholes Ba and Bb, with the sampling horizons marked.</p> Full article ">Figure 4
<p>Diagram of the settlement curve acquisition and prediction method in this study.</p> Full article ">Figure 5
<p>Time-continuous settlement curve in Ra and Rb during the monitoring period of 22 March 2015 to 2 December 2019.</p> Full article ">Figure 6
<p>The predicted settlement curves derived from the three-point modified exponential (TME) method (<span class="html-italic">t</span><sub>1</sub> = 0; prediction models in Ra and Rb are labeled as Rae1 and Rbe1) and hyperbolic (HY) method (<span class="html-italic">t</span><sub>0</sub> = 0; prediction models in Ra and Rb are labeled as Rah1 and Rbh1).</p> Full article ">Figure 7
<p>The geological features of the soil layers around Ba and Bb boreholes. (<b>a</b>) Basic Properties, (<b>b</b>) compressibility, and (<b>c</b>) permeability. Note that according to the United Soil Classification System, the ②<sub>3</sub> Sandy Silt, ⑤<sub>1-1</sub> Clay, and ⑤<sub>1-2</sub> Silty Clay could also be named as silty sand, clayey silt, and clayey silt, respectively. In order to ensure the consistency of soil layers and to facilitate the description, the soil classification in the results and discussion section still uses the soil type (<a href="#remotesensing-13-02418-f003" class="html-fig">Figure 3</a>) that was determined by the geological age, soil behavior, and physical and mechanical properties.</p> Full article ">Figure 8
<p>The micro-scale pore characteristics of the clay around the Ba (CLBA) and clay around Bb (CLBB). (<b>a</b>) Pore distribution, (<b>b</b>) pore type, and (<b>c</b>) morphological fractal dimension.</p> Full article ">Figure 9
<p>Quantitative description of microstructure for clay around Ba (CLBA) and clay around Bb (CLBB). (<b>a</b>) SEM images of clay in 2000× and micro-scale analysis process; (<b>b</b>) statistical microscopic parameters; distributions of (<b>c</b>) equivalent diameter and (<b>d</b>) directional frequency; (<b>e</b>) diagram of the suborbicular structural unit (SU) with complex shape and clay aggregates in 2000×; diagrams of (<b>f</b>) shape and (<b>g</b>) arrangement characteristics of SUs.</p> Full article ">
21 pages, 7246 KiB  
Article
IMG2nDSM: Height Estimation from Single Airborne RGB Images with Deep Learning
by Savvas Karatsiolis, Andreas Kamilaris and Ian Cole
Remote Sens. 2021, 13(12), 2417; https://doi.org/10.3390/rs13122417 - 21 Jun 2021
Cited by 21 | Viewed by 4932
Abstract
Estimating the height of buildings and vegetation in single aerial images is a challenging problem. A task-focused Deep Learning (DL) model that combines architectural features from successful DL models (U-NET and Residual Networks) and learns the mapping from a single aerial imagery to [...] Read more.
Estimating the height of buildings and vegetation in single aerial images is a challenging problem. A task-focused Deep Learning (DL) model that combines architectural features from successful DL models (U-NET and Residual Networks) and learns the mapping from a single aerial imagery to a normalized Digital Surface Model (nDSM) was proposed. The model was trained on aerial images whose corresponding DSM and Digital Terrain Models (DTM) were available and was then used to infer the nDSM of images with no elevation information. The model was evaluated with a dataset covering a large area of Manchester, UK, as well as the 2018 IEEE GRSS Data Fusion Contest LiDAR dataset. The results suggest that the proposed DL architecture is suitable for the task and surpasses other state-of-the-art DL approaches by a large margin. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Aerial images from the first dataset (left), the corresponding nDSMs in heat map format and their color bars, indicating the color-coding of the nDSM in meters (right). The aerial images on the left of each pair have a size of 4000 × 4000, while the size of the nDSMs is 1000 × 1000. nDSMs are presented at the same size as the aerial images for demonstration reasons. The Figure is best seen in color.</p> Full article ">Figure 2
<p>The DSM (<b>left</b>) and the DTM (<b>right</b>) corresponding to the bottom aerial image of <a href="#remotesensing-13-02417-f001" class="html-fig">Figure 1</a>. The color bar for each heat map indicates the color-coding of the DEMs in meters above sea level. Both heat maps have several undetermined or irrational (extremely high or low) values shown in black color. Notably, some of these unexpected values in the DSM map (<b>left</b>) correspond to a river, which illustrates a well-known problem of LiDAR measurements near highly reflective and refractive surfaces with multiple light paths. Such erroneous values raise significant problems regarding the training of the model. Thus, they are detected during data preprocessing and excluded from the training data (see <a href="#sec2dot4-remotesensing-13-02417" class="html-sec">Section 2.4</a>). They are also excluded from the validation and test data to avoid inaccurate performance evaluations. Overall, these values roughly comprise 10% of the dataset but lead to a larger amount of discarded data, since any candidate patch containing even a pixel of undetermined or irrational value is excluded from the training pipeline. This figure is best seen in color.</p> Full article ">Figure 3
<p>Aerial images from the IEEE GRSS Data Fusion Contest (second dataset), the corresponding nDSMs and the color bars of the heat maps indicating the color-coding in meters. The RGB images on the left of each pair have a size of 5000 × 5000 pixels, while the size of the nDSMs is 500 × 500 pixels. The heat maps are shown as the same size as the aerial images for demonstration reasons. This figure is best seen in color.</p> Full article ">Figure 4
<p>The architecture of the three types of residual blocks used in the proposed models: (<b>a</b>) the typical residual block (RBLK). (<b>b</b>) The down-sampling residual block (DRBLK) uses a stride of two at both the first convolutional layer and the skip connection. (<b>c</b>) The up-sampling residual block (URBLK) uses subpixel upscaling at the first convolution and the skip connection. BN stands for batch normalization [<a href="#B55-remotesensing-13-02417" class="html-bibr">55</a>], PReLU for parametric ReLU and s is the stride of the convolutional layer.</p> Full article ">Figure 5
<p>The architecture of the model trained with the Manchester dataset. All convolutional layers use kernel size 3 and “same” padding. BN represents a Batch Normalization layer and CNT a Concatenation layer.</p> Full article ">Figure 6
<p>The architecture of the model trained with the DFC2018 dataset. Compared to the model trained with the Manchester dataset, the kernel sizes of certain convolutional layers are increased, and their padding is changed from “same” to “valid”, as indicated by the figure notes. Additionally, some convolutional layers are introduced at the end of the model. These modifications aim at applying a reduction factor of 10 between the input and the output of the model to match the resolution ratio between the aerial images and the nDSMs in the DFC2018 dataset. BN represents a Batch Normalization layer and CNT a Concatenation layer.</p> Full article ">Figure 7
<p>Left: RGB images of an area in the test set of the Manchester area dataset. Middle: The ground truth nDSMs. Right: The elevation heat maps as predicted by the model. Note 1 shows cases of spurious points in the ground truth that the model correctly avoids estimating. Note 2 shows occasional inconsistencies in the dataset due to the different acquisition times of the RGB images and the LiDAR measurements. Although these inconsistencies are also evident in the training set, the model is robust to such problematic training instances. Note 3 shows cases where the model produces better-quality maps than the ground truth in terms of the surface smoothness and level of detail, as the LiDAR data contains noisy values.</p> Full article ">Figure 8
<p>Left: RGB images from the DFC2018 test set. Middle: Ground truth nDSMs. Right: Model’s height estimations. Note 1 indicates an area that contains a group of trees and is magnified in <a href="#remotesensing-13-02417-f009" class="html-fig">Figure 9</a> to demonstrate how the model treats vegetation in the RGB images.</p> Full article ">Figure 9
<p>Magnification of the noted region (Note 1) in <a href="#remotesensing-13-02417-f008" class="html-fig">Figure 8</a>. Left: The magnified RGB image. Middle: The ground truth nDSM. Right: Model output. The model consistently overestimates the foliage volume by filling the spaces between foliage with similar values to the neighboring estimations.</p> Full article ">Figure 10
<p>Using a sliding window to investigate whether the model uses shadows for object height estimation. Each test case is presented in pairs of consecutive rows, the upper row showing the RGB image with the position of the sliding masking window (black square) and the lower row the prediction at the model output. The artificial shadow implied by the square black box influences the height estimation of the buildings close to the shadow by increasing their height predictions (the values on the predicted map corresponding to the buildings that are close to the shadow are seen to be brighter in the image and, thus, higher in value). The estimated heights of buildings that are not near the implied shadow are not affected. The artificial shadow causes the model to predict a higher elevation for buildings that are in the shadow’s proximity. This figure is better seen in color.</p> Full article ">Figure 11
<p>Sample failed cases where the model misses the presence of an object completely. The cases are magnified regions from the second RGB image (second row) of <a href="#remotesensing-13-02417-f008" class="html-fig">Figure 8</a>. The top-left image shows a very high pole standing on a highway (on the left of the train wagons) with a height of 30 m (according to its LiDAR measurement). Despite the pole’s long shadow, the model does not detect it. The bottom-left magnified region contains a tall electric energy transmission tower (close and on the right of the train wagons) that is also not detected by the model.</p> Full article ">
17 pages, 25245 KiB  
Article
Potential Land Use Conflict Identification Based on Improved Multi-Objective Suitability Evaluation
by Wenli Jing, Kanhua Yu, Lian Wu and Pingping Luo
Remote Sens. 2021, 13(12), 2416; https://doi.org/10.3390/rs13122416 - 20 Jun 2021
Cited by 28 | Viewed by 4155
Abstract
Accurately identifying potential land use conflicts (LUCs) is critical for alleviating the ever-intensifying contradictions between humans and nature. The previous studies using the method of suitability analysis did not take full advantage of the current land use and multi-function characteristics of land resources. [...] Read more.
Accurately identifying potential land use conflicts (LUCs) is critical for alleviating the ever-intensifying contradictions between humans and nature. The previous studies using the method of suitability analysis did not take full advantage of the current land use and multi-function characteristics of land resources. In this study, an improved model of suitability analysis was realized. In order to explore the LUCs status, including the types, intensity and distribution, a multi-objective suitability evaluation model was constructed from the perspective of production-living-ecological functions. And it was applied to Hengkou District, a typical region of the Qin-Ba mountainous area in the central part of China. The results show that the suitability distribution of living- production-ecological functions vary widely from the center to the periphery with altitude in Hengkou District; 22.03% of the land is at a risk of land use conflict. Among them, the high potential conflict areas account for 55.32%, and the conflicts between production and ecological lands (L2P1E1, L3P1E1) are the largest, which are located at the fringe of the central urban and ecologically dominant area. Therefore, it is necessary to adopt effective strategies to achieve a balance between the differential demands of land use. This research could better reflect the true situation of land use in ecologically sensitive mountainous areas and would provide theoretical and methodological support for the identification and prevention of potential LUCs. Full article
(This article belongs to the Special Issue Remote Sensing Approaches to Landscape Analysis of Urban and Peri-Urban Environments)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Research location and land use status map for 2017. (<b>a</b>) Qin-Ba mountainous area in China; (<b>b</b>) Elevation analysis of Qin-Ba mountainous area and the site of Hengkou District; (<b>c</b>) Land use of Hengkou District.</p> Full article ">Figure 2
<p>The flowchart of analytical steps of the research.</p> Full article ">Figure 3
<p>Empirical model for identification of potential land use conflicts.</p> Full article ">Figure 4
<p>Collection of single-factor spatial analysis results. (<b>a</b>) Slope aspect; (<b>b</b>) Slope; (<b>c</b>) Altitude; (<b>d</b>) NDVI; (<b>e</b>) Soil texture; (<b>f</b>) Distance of rivers; (<b>g</b>) Distance of natural disasters; (<b>h</b>) Ecological protection line; (<b>i</b>) Boundary of construction land; (<b>j</b>) Boundary of basic farmland; (<b>k</b>) Grain yield per unit area; (<b>l</b>) Per capita cultivated land area; (<b>m</b>) Population density; (<b>n</b>) Per capita net income; (<b>o</b>) Distance from main traffic routes; (<b>p</b>) Classification of ecological land; (<b>q</b>) Classification of production land; (<b>r</b>) Classification of living land; (<b>s</b>) Distance from main service facilities; (<b>t</b>) Distance from the town.</p> Full article ">Figure 5
<p>Unit frequency distribution histogram of Living-Production-Ecological suitability score. (<b>a</b>) Unit frequency distribution histogram of living suitability score; (<b>b</b>) Unit frequency distribution histogram of production suitability score; (<b>c</b>) Unit frequency distribution histogram of ecological suitability score.</p> Full article ">Figure 6
<p>Distribution map of Living-Production-Ecological suitability in Hengkou District. (<b>a</b>) Map of living suitability; (<b>b</b>) Map of production suitability; (<b>c</b>) Map of ecological suitability.</p> Full article ">Figure 7
<p>Distribution of dominant and potential conflict areas in Hengkou District. (<b>a</b>) Map of dominant and potential conflict areas; (<b>b</b>) Map of dominant areas; (<b>c</b>) Map of potential conflict areas.</p> Full article ">Figure 8
<p>Distribution of potential conflict areas based on type and intensity classification in Hengkou District. (<b>a</b>) Map of high conflict areas; (<b>b</b>) Map of moderate potential conflict areas; (<b>c</b>) Map of low potential conflict areas.</p> Full article ">
22 pages, 7178 KiB  
Article
Investigating the Temporal and Spatial Dynamics of Human Development Index: A Comparative Study on Countries and Regions in the Eastern Hemisphere from the Perspective of Evolution
by Hanwei Liang, Na Li, Ji Han, Xin Bian, Huaixia Xia and Liang Dong
Remote Sens. 2021, 13(12), 2415; https://doi.org/10.3390/rs13122415 - 20 Jun 2021
Viewed by 3743
Abstract
The Human Development Index (HDI) is a prevailing indicator to present the status and trend of sustainability of nations, hereby offers a valuable measurement on the Sustainable Development Goals (SDGs). Revealing the dynamics of the HDI of the Eastern Hemisphere countries is vital [...] Read more.
The Human Development Index (HDI) is a prevailing indicator to present the status and trend of sustainability of nations, hereby offers a valuable measurement on the Sustainable Development Goals (SDGs). Revealing the dynamics of the HDI of the Eastern Hemisphere countries is vital for measurement and evaluation of the human development process and revealing the spatial disparities and evolutionary characteristics of human development. However, the statistical data-based HDI, which is currently widely applied, has defects in terms of data availability and inconsistent statistical caliber. To tackle such an existing gap, we applied nighttime lights (NTL) data to reconstruct new HDI indicators named HDINTL and quantify the HDINTL at multispatial scales of Eastern Hemisphere countries during 1992–2013. Results showed that South Central Asia countries had the smallest discrepancies in HDINTL, while the largest was found in North Africa. The national-level HDINTL values in the Eastern Hemisphere ranged between 0.138 and 0.947 during 1992–2013. At the subnational scale, the distribution pattern of HDINTL was spatially clustered based on the results of spatial autocorrelation analysis. The evolutionary trajectory of subnational level HDINTL exhibited a decreasing and then increasing trend along the northwest to the southeast direction of Eastern Hemisphere. At the pixel scale, 93.52% of the grids showed an increasing trend in HDINTL, especially in the urban agglomerations of China and India. These results are essential for the ever-improvement of policy making to reduce HDI’s regional disparity and promote the continuous development of humankind’s living qualities. This study offers an improved HDI accounting method. It expects to extend the channel of HDI application, e.g., potential integration with environmental, physical, and socioeconomic data where the NTL data could present as well. Full article
(This article belongs to the Special Issue Time Series Analysis in Remote Sensing: Algorithm Development and Applications)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Theoretical framework.</p> Full article ">Figure 2
<p>Spatial distribution of countries and regions in the Eastern Hemisphere.</p> Full article ">Figure 3
<p>Loglog linear correlation between the total nighttime light (NTL) and gross national product (GNP) for each subnational unit in the Eastern Hemisphere in (<b>a</b>) 1992, (<b>b</b>) 1999, (<b>c</b>) 2006, and (<b>d</b>) 2013. Note that the different colored scatters are to distinguish the subnational units located in different regions.</p> Full article ">Figure 4
<p>Correlation analysis between the subnational HDI<sub>NTL</sub> (SHDI<sub>NTL</sub>) and the reported Subnational Human Development Index (SHDI) in (<b>a</b>) 1992, (<b>b</b>) 1999, (<b>c</b>) 2006, and (<b>d</b>) 2013. Note that the reported SHDI refers to the data reported by the Subnational Human Development Database [<a href="#B40-remotesensing-13-02415" class="html-bibr">40</a>].</p> Full article ">Figure 5
<p>Range of the HDI<sub>NTL</sub> at national scale in the Eastern Hemisphere, 1992–2013. Note that the number on the y-axis represents the national HDI<sub>NTL</sub> value. The red line is the average of national HDI<sub>NTL</sub> for each region.</p> Full article ">Figure 6
<p>(<b>a</b>) The average value of HDI<sub>NTL</sub> at national scale in the Eastern Hemisphere, 1992–2013. (<b>b</b>) The spatial distribution map of HDI<sub>NTL</sub> at national scale in 2013.</p> Full article ">Figure 7
<p>Hotspot analysis result (Getis-Ord<math display="inline"> <semantics> <mrow> <msubsup> <mi>G</mi> <mi>i</mi> <mo>∗</mo> </msubsup> </mrow> </semantics> </math>statistic) of HDI<sub>NTL</sub> at subnational scale in the Eastern Hemisphere in (<b>a</b>) 1992, (<b>b</b>) 1999, (<b>c</b>) 2006, and (<b>d</b>) 2013.</p> Full article ">Figure 8
<p>Categories of HDI<sub>NTL</sub> at subnational scale in the Eastern Hemisphere in (<b>a</b>) 1992, (<b>b</b>) 1998, (<b>c</b>) 2003, (<b>d</b>) 2008, and (<b>e</b>) 2013. Note that there are 1076 subnational units for which SHDI<sub>NTL</sub> data are available.</p> Full article ">Figure 9
<p>HDI<sub>NTL</sub> evolution process at subnational scale in the Eastern Hemisphere, 1992–2013. Note that the number on the y-axis represents the human development level: 1 indicates the Low-level, 2 indicates the Medium-level, 3 indicates the High-level, and 4 indicates the Very-high-level.</p> Full article ">Figure 10
<p>Spatial distribution map of HDI<sub>NTL</sub> at pixel scale in the Eastern Hemisphere in (<b>a</b>) 1992 and (<b>b</b>) 2013. Note that NoData areas are shown in gray.</p> Full article ">Figure 11
<p>Spatial variation map of HDI<sub>NTL</sub> at pixel scale in the Eastern Hemisphere, 1992–2013. Note that the NoData areas are shown in gray.</p> Full article ">Figure 12
<p>Evolutionary trend of HDI<sub>NTL</sub> at pixel scale in the Eastern Hemisphere, 1992–2013. Note that the bottom three maps highlight four representative urban agglomerations in pink. From left to right, they are Mumbai Urban Agglomerations in India, Guangdong-Hong Kong-Macao Greater Bay Area and Yangtze River Delta Urban Agglomerations in China, Tokyo Metropolitan Area in Japan. NoData areas are shown in white.</p> Full article ">
18 pages, 3715 KiB  
Article
A Comprehensive Evaluation of Five Evapotranspiration Datasets Based on Ground and GRACE Satellite Observations: Implications for Improvement of Evapotranspiration Retrieval Algorithm
by Lijun Chao, Ke Zhang, Jingfeng Wang, Jin Feng and Mengjie Zhang
Remote Sens. 2021, 13(12), 2414; https://doi.org/10.3390/rs13122414 - 20 Jun 2021
Cited by 60 | Viewed by 5723
Abstract
Evapotranspiration (ET) is a vital part of the hydrological cycle and the water–energy balance. To explore the characteristics of five typical remote sensing evapotranspiration datasets and provide guidance for algorithm development, we used reconstructed evapotranspiration (Recon) data based on ground and GRACE satellite [...] Read more.
Evapotranspiration (ET) is a vital part of the hydrological cycle and the water–energy balance. To explore the characteristics of five typical remote sensing evapotranspiration datasets and provide guidance for algorithm development, we used reconstructed evapotranspiration (Recon) data based on ground and GRACE satellite observations as a benchmark and evaluated five remote sensing datasets for 592 watersheds across the continental United States. The Global Land Evaporation Amsterdam Model (GLEAM) dataset (with bias and RMSE values of 23.18 mm/year and 106.10 mm/year, respectively), process-based land surface evapotranspiration/heat flux (P-LSH) dataset (bias = 22.94 mm/year and RMSE = 114.44 mm/year) and the Penman–Monteith–Leuning (PML) algorithm generated ET dataset (bias = −17.73 mm/year and RMSE = 108.97 mm/year) showed the better performance on a yearly scale, followed by the model tree ensemble (MTE) dataset (bias = 99.45 mm/year and RMSE = 141.32 mm/year) and the moderate-resolution imaging spectroradiometer (MODIS) dataset (bias = −106.71 mm/year and RMSE = 158.90 mm/year). The P-LSH dataset outperformed the other four ET datasets on a seasonal scale, especially from March to August. Both PML and MTE showed better overall accuracy and could accurately capture the spatial variability of evapotranspiration in arid regions. The P-LSH and GLEAM products were consistent with the Recon data in middle-value section. MODIS and MTE had larger bias and RMSE values on a yearly scale, whereby the MODIS and MTE datasets tended to underestimate and overestimate ET values in all the sections, respectively. In the future, the aim should be to reduce bias in the MODIS and MTE algorithms and further improve seasonality of the ET estimation in the GLEAM algorithm, while the estimation accuracy of the P-LSH and MODIS algorithms should be improved in arid regions. Our analysis suggests that combining artificial intelligence algorithms or data-driven algorithms and physical process algorithms will further improve the accuracy of ET estimation algorithms and the quality of ET datasets, as well as enhancing their capacity to be applied in different climate regions. Full article
(This article belongs to the Section Environmental Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Map of the study area and location of 592 watersheds.</p> Full article ">Figure 2
<p>Yearly scale comparison of (<b>a</b>) P-LSH, (<b>b</b>) PML, (<b>c</b>) MODIS, (<b>d</b>) MTE, and (<b>e</b>) GLEAM data to Recon data.</p> Full article ">Figure 3
<p>Maps of multiyear (2003–2008) average annual ET in this study: (<b>a</b>) Recon, (<b>b</b>) P-LSH, (<b>c</b>) PML, (<b>d</b>) MODIS, (<b>e</b>) MTE, and (<b>f</b>) GLEAM .</p> Full article ">Figure 4
<p>Box plot of multiyear (2003–2008) average annual ET in this study.</p> Full article ">Figure 5
<p>The spatial distributions of multiyear (2003–2008) mean seasonal ET for five datasets in CONUS: (<b>a</b>–<b>d</b>) spring (March, April, and May (MAM)), summer (June, July, and August (JJA)), autumn (Septemeber, October, and November (SON)), and winter (December, January, and Feburary (DJF)) average ET of Recon; (<b>e</b>–<b>h</b>) MAM, JJA, SON, and DJF average ET of P-LSH; (<b>i</b>–<b>l</b>) MAM, JJA, SON, and DJF average ET of PML; (<b>m</b>–<b>p</b>) MAM, JJA, SON, and DJF average ET of MODIS; (<b>q</b>–<b>t</b>) MAM, JJA, SON, and DJF average ET of MTE; and (<b>u</b>–<b>x</b>) MAM, JJA, SON, and DJF average ET of GLEAM.</p> Full article ">Figure 6
<p>Comparison of regional average ET derived from Recon, P-LSH, PML, MODIS, MTE, and GLEAM.</p> Full article ">Figure 7
<p>Taylor diagram of monthly P-LSH, PML, MODIS, MTE, and GLEAM data based on comparison with monthly Recon for all grid cells from 2003 to 2008.</p> Full article ">Figure 8
<p>Spatial maps of correlation coefficients (<b>a</b>–<b>e</b>) between monthly Recon and five monthly remote sensing ET datasets and their significance levels (<b>f</b>–<b>j</b>) in this study.</p> Full article ">Figure 9
<p>Quantile–quantile plot of Recon data vs. five remote sensing ET datasets for all watersheds.</p> Full article ">
16 pages, 3186 KiB  
Article
Preliminary Significant Wave Height Retrieval from Interferometric Imaging Radar Altimeter Aboard the Chinese Tiangong-2 Space Laboratory
by Lin Ren, Jingsong Yang, Xiao Dong, Yongjun Jia and Yunhua Zhang
Remote Sens. 2021, 13(12), 2413; https://doi.org/10.3390/rs13122413 - 20 Jun 2021
Cited by 9 | Viewed by 2379
Abstract
The interferometric imaging radar altimeter (InIRA) aboard the Chinese Tiangong-2 space laboratory is the first spaceborne imaging radar working at low incidence angles. This study focuses on the retrieval of significant wave heights (SWHs) from InIRA data. The retrieved SWHs can be used [...] Read more.
The interferometric imaging radar altimeter (InIRA) aboard the Chinese Tiangong-2 space laboratory is the first spaceborne imaging radar working at low incidence angles. This study focuses on the retrieval of significant wave heights (SWHs) from InIRA data. The retrieved SWHs can be used for correcting the sea state bias of InIRA-derived sea surface heights and can supplement SWH products from other spaceborne sensors. First, we analyzed tilt, range bunching and velocity bunching wave modulations at low incidence angles, and we found clear dependencies between the SWH and two defined factors, range and azimuth integration, for ocean waves in the range and azimuth directions, respectively. These dependencies were further confirmed using InIRA measurements and collocated WaveWatch III (WW3) data. Then, an empirical orthogonal SWH model using the range and azimuth integration factors as model inputs was proposed. The model was segmented by the incidence angle, and the model coefficients were estimated by fitting the collocation at each incidence angle bin. Finally, the SWHs were retrieved from InIRA data using the proposed model. The retrievals were validated using both WW3 and altimeter (JASON2, JASON3, SARAL, and HY2A) SWHs. The validation with WW3 data shows a root mean square error (RMSE) of 0.43 m, while the average RMSE with all traditional altimeter data is 0.48 m. This indicates that the InIRA can be used to measure SWHs. Full article
(This article belongs to the Special Issue Synthetic Aperture Radar Observations of Marine Coastal Environments-II)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location map of InIRA data and collocated traditional altimeter data comprising JASON2, JASON3, SARAL and HY2A. The blue lines denote the InIRA data track, and the gold, green, purple and red “+” signs, respectively, denote the collocated JASON2, JASON3, SARAL, and HY2A positions with the InIRA.</p> Full article ">Figure 2
<p>Two InIRA data cases: (<b>a</b>) NRCS image for swell and (<b>b</b>) image spectra extracted from (<b>a</b>); (<b>c</b>) NRCS image for crossing wind waves and (<b>d</b>) image spectra extracted from (<b>c</b>).</p> Full article ">Figure 3
<p>InIRA data dependence analysis: (<b>a</b>) SWH dependence on the range integration factor for data close to the range direction; (<b>b</b>) SWH dependence on the azimuth integration factor for data close to the azimuth direction; (<b>c</b>) same as (<b>a</b>) but for data close to the azimuth direction; (<b>d</b>) same as (<b>b</b>) but for data close to the range direction; (<b>e</b>) range integration factor dependence on the incidence angle for data close to the range direction; (<b>f</b>) the azimuth integration factor dependence on the incidence angle for data close to the azimuth direction. The green points represent the data close to the range direction, while the blue points are for the azimuth direction.</p> Full article ">Figure 4
<p>SWH data from (<b>a</b>) InIRA, (<b>b</b>) SARAL, and (<b>c</b>) HY2A. The black lines are plotted to find the intersection positions.</p> Full article ">Figure 5
<p>Comparisons between SWHs from InIRA retrievals and collocated (<b>a</b>) SARAL and (<b>b</b>) HY2A data.</p> Full article ">Figure 6
<p>SWH comparison between InIRA retrievals and collocated WW3 data.</p> Full article ">Figure 7
<p>SWH accuracy trends in terms of (<b>a</b>) BIAS with incidence angle; (<b>b</b>) RMSE with incidence angle; (<b>c</b>) BIAS with relative wave direction; and (<b>d</b>) RMSE with relative wave direction.</p> Full article ">Figure 8
<p>Comparisons of SWH between InIRA retrievals and collocated (<b>a</b>) JASON2; (<b>b</b>) JASON3; (<b>c</b>) SARAL; and (<b>d</b>) HY2A data.</p> Full article ">
20 pages, 4583 KiB  
Article
Measurements of Rainfall Rate, Drop Size Distribution, and Variability at Middle and Higher Latitudes: Application to the Combined DPR-GMI Algorithm
by Viswanathan Bringi, Mircea Grecu, Alain Protat, Merhala Thurai and Christian Klepp
Remote Sens. 2021, 13(12), 2412; https://doi.org/10.3390/rs13122412 - 20 Jun 2021
Cited by 7 | Viewed by 3058
Abstract
The Global Precipitation Measurement mission is a major U.S.–Japan joint mission to understand the physics of the Earth’s global precipitation as a key component of its weather, climate, and hydrological systems. The core satellite carries a dual-precipitation radar and an advanced microwave imager [...] Read more.
The Global Precipitation Measurement mission is a major U.S.–Japan joint mission to understand the physics of the Earth’s global precipitation as a key component of its weather, climate, and hydrological systems. The core satellite carries a dual-precipitation radar and an advanced microwave imager which provide measurements to retrieve the drop size distribution (DSD) and rain rates using a Combined Radar-Radiometer Algorithm (CORRA). Our objective is to validate key assumptions and parameterizations in CORRA and enable improved estimation of precipitation products, especially in the middle-to-higher latitudes in both hemispheres. The DSD parameters and statistical relationships between DSD parameters and radar measurements are a central part of the rainfall retrieval algorithm, which is complicated by regimes where DSD measurements are abysmally sparse (over the open ocean). In view of this, we have assembled optical disdrometer datasets gathered by research vessels, ground stations, and aircrafts to simulate radar observables and validate the scattering lookup tables used in CORRA. The joint use of all DSD datasets spans a large range of drop concentrations and characteristic drop diameters. The scaling normalization of DSDs defines an intercept parameter NW, which normalizes the concentrations, and a scaling diameter Dm, which compresses or stretches the diameter coordinate axis. A major finding of this study is that a single relationship between NW and Dm, on average, unifies all datasets included, from stratocumulus to heavier rainfall regimes. A comparison with the NW–Dm relation used as a constraint in versions 6 and 7 of CORRA highlights the scope for improvement of rainfall retrievals for small drops (Dm < 1 mm) and large drops (Dm > 2 mm). The normalized specific attenuation–reflectivity relationships used in the combined algorithm are also found to match well the equivalent relationships derived using DSDs from the three datasets, suggesting that the currently assumed lookup tables are not a major source of uncertainty in the combined algorithm rainfall estimates. Full article
(This article belongs to the Special Issue Remote Sensing of Precipitation at the Mid- to High-Latitudes)
►▼ Show Figures

Figure 1

Figure 1
<p>Locations of sites where MPS and 2DVD were deployed for Ground Validation. The purple rectangle depicts the region penetrated by the C-130. The red lines depict the tracks of the R/V Investigator in 2016 and 2018.</p> Full article ">Figure 2
<p>Density plot of h(x) from stratocumulus drizzle with number of occurrences given by the color bar in contours of log N. The GXY-HSV data are pooled together for D<sub>m</sub> &lt; 0.5 mm and shown as black points. The dashed yellow line is the h(x) for the gamma model with μ = 3.</p> Full article ">Figure 3
<p>Relative frequency histograms of (<b>a</b>) D<sub>m</sub> from stratocumulus drizzle, (<b>b</b>) same as (<b>a</b>) but from OceanRain and pooled DSDs from GXY-HSV, and (<b>c</b>) histograms of log<sub>10</sub> (N<sub>W</sub>) from OceanRain and pooled DSDs and from stratocumulus drizzle (red line).</p> Full article ">Figure 4
<p>Scatterplot of the reflectivity at the Ku-band versus D<sub>m</sub> from stratocumulus drizzle (density-plot in color) and from all DSDs from GXY-HSV (blue points).</p> Full article ">Figure 5
<p>Scatterplot of log<sub>10</sub> (N<sub>W</sub>) versus D<sub>m</sub> from three regimes described in the legend. The contoured density plot (refer to color bar) is from OceanRain. The mean fit to OceanRain is N<sub>W</sub> = 6383.8 D<sub>m</sub><sup>−3.19</sup>, where N<sub>W</sub> is in mm<sup>−1</sup> m<sup>−3</sup> shown as red-filled black squares. The GXY- and HSV-based points are shown in light red and those from stratocumulus DSDs are shown in magenta.</p> Full article ">Figure 6
<p>The coefficient α versus shape factor μ for a theoretical gamma with fall speed power law v(D) in m s<sup>−1</sup>, D in mm, R in mm h<sup>−1</sup>, and N<sub>W</sub> in mm<sup>−1</sup> m<sup>−3</sup>. Values of α are given within textboxes for μ = −2 and μ = 10.</p> Full article ">Figure 7
<p>Rain rate (in mm h<sup>−1</sup>) normalized by N<sub>W</sub> (in mm<sup>−1</sup> m<sup>−3</sup>) versus D<sub>m</sub> (in mm). We compare random gamma DSDs with the theoretical result in Equation (1) with μ = 0.</p> Full article ">Figure 8
<p>As in <a href="#remotesensing-13-02412-f007" class="html-fig">Figure 7</a>, except DSDs from stratocumulus drizzle, OceanRain, and GXY-HSV.</p> Full article ">Figure 9
<p>Ku-band-specific attenuation (k<sub>ku</sub> in dB km<sup>−1</sup>) normalized by Ku-band reflectivity (Z<sub>ku</sub> in mm<sup>6</sup> m<sup>−3</sup>) versus D<sub>m</sub> (in mm), from OceanRain DSDs, GXY-HSV DSDs, DSDs from the outer bands of Tropical Storm Irma and Tropical Depression Nate, as well as category-1 Hurricane Dorian, compared with those from the combined algorithm (CMB).</p> Full article ">Figure 10
<p>Variation of the rain rate (R in mm h<sup>−1</sup>) normalized by N<sub>W</sub> (in mm<sup>−1</sup> m<sup>−3</sup>) versus Ku-band reflectivity (Z<sub>ku</sub>, in mm<sup>6</sup> m<sup>−3</sup>). See legend, which lists the regimes.</p> Full article ">Figure 11
<p>As in <a href="#remotesensing-13-02412-f010" class="html-fig">Figure 10</a>, but for Ka-band.</p> Full article ">Figure 12
<p>Variation of Ku-band specific attenuation (k<sub>ku</sub> in dB km<sup>−1</sup>) normalized by N<sub>W</sub> (mm<sup>−1</sup> m<sup>−3</sup>) versus Ku-band reflectivity (Z<sub>ku</sub>, in mm<sup>6</sup> m<sup>−3</sup>) normalized by N<sub>W</sub>. See legend, which lists the regimes.</p> Full article ">Figure 13
<p>As in <a href="#remotesensing-13-02412-f012" class="html-fig">Figure 12</a> but for Ka-band.</p> Full article ">Figure 14
<p>Two-dimensional histogram (or density plot) of joint (N<sub>w</sub>, D<sub>m</sub>) variables derived using version 6 (V6; panel <b>a</b>) and version 7 (V7; panel <b>b</b>) of the combined algorithm. Superimposed are the mean N<sub>W</sub>–D<sub>m</sub> derived from the Validation Network (VN) and used to constrain the retrievals in V6 and V7 (symbolized with triangles) and the fit to OceanRain N<sub>W</sub>–D<sub>m</sub> data (symbolized with squares).</p> Full article ">Figure A1
<p>(<b>a</b>) Histograms of <math display="inline"><semantics> <mrow> <mo>∆</mo> <msub> <mrow> <mrow> <mi>log</mi> </mrow> </mrow> <mrow> <mn>10</mn> </mrow> </msub> <msub> <mi mathvariant="normal">N</mi> <mi mathvariant="normal">W</mi> </msub> <mo> </mo> </mrow> </semantics></math> for the GXY-HSV (green) and OceanRain (rain) datasets, and their fitted curves; (<b>b</b>) the same as (<b>a</b>) but with drizzle histograms included (blue).</p> Full article ">
14 pages, 2455 KiB  
Technical Note
Detection of Changes in Arable Chernozemic Soil Health Based on Landsat TM Archive Data
by Igor Savin, Elena Prudnikova, Yury Chendev, Anastasia Bek, Dmitry Kucher and Petr Dokukin
Remote Sens. 2021, 13(12), 2411; https://doi.org/10.3390/rs13122411 - 19 Jun 2021
Cited by 4 | Viewed by 2235
Abstract
When soils are used for a long period of time as arable land, their properties change. This can lead to soil degradation and loss of fertility, as well as other important soil biosphere functions. Obtaining data on the trends in arable soil conditions [...] Read more.
When soils are used for a long period of time as arable land, their properties change. This can lead to soil degradation and loss of fertility, as well as other important soil biosphere functions. Obtaining data on the trends in arable soil conditions over large areas using traditional field survey methods is expensive and time-consuming. Currently, there are large archives of satellite data that can be used to monitor the status of arable soils. The analysis of changes in the color of the surface of arable chernozem soils of the Belgorod region, for the period from 1985 to the present, has been carried out based on the analysis of Landsat TM5 satellite data and information about the spectral reflectance of the soils of the region. It is found that, on most parts of arable lands of the region, the color of the soil surface has not changed significantly since 1985. Color changes were revealed on 11% of the analyzed area. The greatest changes are connected with the humus content and moisture content of soils. The three most probable reasons for the change of humus content in an arable horizon of soils are as follows: the dehumidification of soils during plowing; the reduction of the humus content due to water erosion; and the increase in humus content due to changes in the land-use system of the region in recent years. The change in soil moisture regime has mainly been found in arable lands in river valleys, most likely conditioned by the natural evolution of soils. Trends of increasing soil moisture are prevalent. The revealed regularities testify to the high stability of arable soils in the region during the last few decades. Full article
(This article belongs to the Special Issue Land and Soil Health Assessment and Monitoring Based on Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Geographical position of the study area (red oval).</p> Full article ">Figure 2
<p>Flowchart of the investigation.</p> Full article ">Figure 3
<p>Examples of Landsat images (2,3,4 bands composite) for the region of research, <b>left</b>—11 May 1985, <b>right</b>—3 May 2011.</p> Full article ">Figure 4
<p>Expert-based model of soil spectral reflectance in Landsat TM5 bands. 1—average spectral curve for dominant nondegraded soils of the region; 2—soil with higher humus content than dominant; 3—soil with lower humus content than dominant; 4—soil with higher carbonates content than dominant; 5—soil with higher moisture than dominant.</p> Full article ">Figure 5
<p>Map of soil property changes of the studied region from 1985 to 2010.</p> Full article ">
23 pages, 6023 KiB  
Article
Spatial Patterns in Actual Evapotranspiration Climatologies for Europe
by Simon Stisen, Mohsen Soltani, Gorka Mendiguren, Henrik Langkilde, Monica Garcia and Julian Koch
Remote Sens. 2021, 13(12), 2410; https://doi.org/10.3390/rs13122410 - 19 Jun 2021
Cited by 17 | Viewed by 3989
Abstract
Spatial patterns in long-term average evapotranspiration (ET) represent a unique source of information for evaluating the spatial pattern performance of distributed hydrological models on a river basin to continental scale. This kind of model evaluation is getting increased attention, acknowledging the shortcomings of [...] Read more.
Spatial patterns in long-term average evapotranspiration (ET) represent a unique source of information for evaluating the spatial pattern performance of distributed hydrological models on a river basin to continental scale. This kind of model evaluation is getting increased attention, acknowledging the shortcomings of traditional aggregated or timeseries-based evaluations. A variety of satellite remote sensing (RS)-based ET estimates exist, covering a range of methods and resolutions. There is, therefore, a need to evaluate these estimates, not only in terms of temporal performance and similarity, but also in terms of long-term spatial patterns. The current study evaluates four RS-ET estimates at moderate resolution with respect to spatial patterns in comparison to two alternative continental-scale gridded ET estimates (water-balance ET and Budyko). To increase comparability, an empirical correction factor between clear sky and all-weather ET, based on eddy covariance data, is derived, which could be suitable for simple corrections of clear sky estimates. Three RS-ET estimates (MODIS16, TSEB and PT-JPL) and the Budyko method generally display similar spatial patterns both across the European domain (mean SPAEF = 0.41, range 0.25–0.61) and within river basins (mean SPAEF range 0.19–0.38), although the pattern similarity within river basins varies significantly across basins. In contrast, the WB-ET and PML_V2 produced very different spatial patterns. The similarity between different methods ranging over different combinations of water, energy, vegetation and land surface temperature constraints suggests that robust spatial patterns of ET can be achieved by combining several methods. Full article
(This article belongs to the Special Issue Remote Sensing for Distributed Hydrologic Models: New Satellite Data, Model Parametrization and Spatial Metrics to Calibrate and Evaluate Models)
►▼ Show Figures

Figure 1

Figure 1
<p>The study area showing the distributions of 23 selected catchments and 30 FLUXNET sites with the underlying topography across Europe.</p> Full article ">Figure 2
<p>The water-balance ET approach components at a 25 km spatial resolution: (<b>a</b>) E-Obs precipitation and (<b>b</b>) E-Run surface runoff, both for the period of 2002–2014.</p> Full article ">Figure 3
<p>The Budyko curve representing the energy-limit (ET = PET) and the water-limit (ET = P) lines.</p> Full article ">Figure 4
<p>The Budyko ET approach components at a 25 km spatial resolution: (<b>a</b>) potential evapotranspiration and (<b>b</b>) distributed n parameter values derived from NDVI, both for the period of 2002–2012.</p> Full article ">Figure 5
<p>The long-term monthly mean (2002–2014) latent heat flux (LE) during all days and sunny days for 30 FLUXNET stations across Europe. See <a href="#remotesensing-13-02410-f001" class="html-fig">Figure 1</a> for the spatial distribution of the EC sites.</p> Full article ">Figure 6
<p>The grid-based (25 km) annual mean benchmark ET spatial patterns over the period of 2002–2014: (<b>a</b>) water-balance WB-ET and (<b>b</b>) Budyko ET. The catchments boundaries (N: 23) are overlaid on the ET maps. See <a href="#remotesensing-13-02410-f001" class="html-fig">Figure 1</a> for detailed descriptions.</p> Full article ">Figure 7
<p>The grid-based (1 km) annual mean RS-ET spatial patterns over the period of 2002–2014: (<b>a</b>) MOD16, (<b>b</b>) PML_V2, (<b>c</b>) PT-JPL, and (<b>d</b>) TSEB. The catchments boundaries (N: 23) are overlaid on the RS-ET maps. See <a href="#remotesensing-13-02410-f001" class="html-fig">Figure 1</a> for detailed descriptions.</p> Full article ">Figure 8
<p>The catchment-based SPAEF calculation between the RS-ET products with a 1 km spatial resolution over the period of 2002–2014. The catchments are shown in y-axis (rows) and the RS-ET datasets are displayed in x-axis (columns).</p> Full article ">Figure 9
<p>The normalized (each ET map was divided by its mean value) catchment-based annual mean ET spatial patterns with a 25 km spatial resolution over the period of 2002–2014: (<b>a</b>) WB-ET, (<b>b</b>) Budyko ET, (<b>c</b>) MODIS16 ET, (<b>d</b>) PML_V2 ET, (<b>e</b>) PT-JPL ET and (<b>f</b>) TSEB ET.</p> Full article ">Figure 10
<p>(<b>a</b>) Evaluation of the long-term (2002–2014) estimated WB-ET and derived RS-ET products using the Budyko curve; (<b>b</b>) geographical location of the catchments corresponding to the color of points in panels <span class="html-italic">a</span>. The red (blue) box represents catchments with the highest (lowest) aridity index, which fall in the water-limited (energy-limited) environment.</p> Full article ">Figure 11
<p>The empirical Copula densities (1 km grid at full-domain) amongst the satellite RS-ET products. (<b>a</b>–<b>c</b>) MODIS16 vs. other RS-ET estimates (<b>d</b>,<b>e</b>) PML_V2 vs. other RS-ET estimates, and (<b>f</b>) PTJPL vs. TSEB. The sample size is 2,247,522 data tuples for the ET datasets.</p> Full article ">Figure 12
<p>The empirical Copula densities (25 km grid at catchment-scale): (<b>a</b>–<b>d</b>) WB-ET and (<b>e</b>–<b>h</b>) Budyko ET against RS-ET datasets, and (<b>i</b>) WB-ET against Budyko ET. The sample size is 3413 data tuples for the ET datasets.</p> Full article ">Figure 13
<p>The hierarchical cluster analysis for the annual mean (2002–2014) normalized ET datasets in a 25 km grid, representing the overall similarity ranking among the ET products for each of the catchments across the study area. See <a href="#remotesensing-13-02410-f001" class="html-fig">Figure 1</a> for the locations of the catchments.</p> Full article ">
22 pages, 4988 KiB  
Article
Spatiotemporal Continuous Impervious Surface Mapping by Fusion of Landsat Time Series Data and Google Earth Imagery
by Rui Chen, Xiaodong Li, Yihang Zhang, Pu Zhou, Yalan Wang, Lingfei Shi, Lai Jiang, Feng Ling and Yun Du
Remote Sens. 2021, 13(12), 2409; https://doi.org/10.3390/rs13122409 - 19 Jun 2021
Cited by 12 | Viewed by 3384
Abstract
The monitoring of impervious surfaces in urban areas using remote sensing with fine spatial and temporal resolutions is crucial for monitoring urban development and environmental changes in urban areas. Spatiotemporal super-resolution mapping (STSRM) fuses fine-spatial-coarse-temporal remote sensing data with coarse-spatial-fine-temporal data, allowing for [...] Read more.
The monitoring of impervious surfaces in urban areas using remote sensing with fine spatial and temporal resolutions is crucial for monitoring urban development and environmental changes in urban areas. Spatiotemporal super-resolution mapping (STSRM) fuses fine-spatial-coarse-temporal remote sensing data with coarse-spatial-fine-temporal data, allowing for urban impervious surface mapping at both fine-spatial and fine-temporal resolutions. The STSRM involves two main steps: unmixing the coarse-spatial-fine-temporal remote sensing data to class fraction images, and downscaling the fraction images to sub-pixel land cover maps. Yet, challenges exist in each step when applying STSRM in mapping impervious surfaces. First, the impervious surfaces have high spectral variability (i.e., high intra-class and low inter-class variability), which impacts the accurate extraction of sub-pixel scale impervious surface fractions. Second, downscaling the fraction images to sub-pixel land cover maps is an ill-posed problem and would bring great uncertainty and error in the predictions. This paper proposed a new Spatiotemporal Continuous Impervious Surface Mapping (STCISM) method to deal with these challenges in fusing Landsat and Google Earth imagery. The STCISM used the Multiple Endmember Spectral Mixture Analysis and the Fisher Discriminant Analysis to minimize the within-class variability and maximize the between-class variability to reduce the spectral unmixing uncertainty. In addition, the STCISM adopted a new temporal consistency check model to incorporate temporal contextual information to reduce the uncertainty in the time-series impervious surface prediction maps. Unlike the traditional temporal consistency check model that assumed the impervious-to-pervious conversion is unlikely to happen, the new model allowed the bidirectional conversions between pervious and impervious surfaces. The temporal consistency check was used as a post-procession method to correct the errors in the prediction maps. The proposed STCISM method was used to predict time-series impervious surface maps at 5 m resolution of Google Earth image at the Landsat frequency. The results showed that the proposed STCISM outperformed the STSRM model without using the temporal consistency check and the STSRM model using the temporal consistency check based on the unidirectional pervious-to-impervious surface conversion rule. Full article
(This article belongs to the Special Issue Time Series Analysis in Remote Sensing: Algorithm Development and Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>The location of the study area. The right image is the Google Earth image of Daqiao New Town. The top-left image is the scaled map of China in the international context, and the bottom-left corner is Wuhan.</p> Full article ">Figure 2
<p>The number of Landsat Thematic Mapper (TM) and Operational Land Imager (OLI) images acquired in each year used in this study.</p> Full article ">Figure 3
<p>The flowchart of the STCISM.</p> Full article ">Figure 4
<p>An example of transformation from (<b>a</b>) a pervious surface (2006/12/17) to (<b>b</b>) an impervious surface (2013/06/13) and (<b>c</b>) revision back to a pervious surface (2020/02/25).</p> Full article ">Figure 5
<p>Schematic diagram of a temporal consistency check model.</p> Full article ">Figure 6
<p>Visual comparison of the Bayesian-STSRM and STCISM predictions. Red and white represent impervious and pervious surfaces, respectively. GE: Google Earth. (<b>A</b>) Buildings and trees (<b>B</b>) Building and soil objects.</p> Full article ">Figure 7
<p>Visual comparison between our impervious surface maps and other products: Google Earth (GE) images acquired on 16 October 2010, and 20 February 2016; for 2010, Spatiotemporal Continuous Impervious Surface Mapping (STCISM) dataset; Global Artificial Impervious Areas (GAIA) dataset, and Normalized Urban Areas Composite Index (NUACI) dataset; and for 2015, STCISM dataset, GAIA dataset, and Multisource, Multitemporal (MSMT) dataset. Red and white represent impervious and pervious surfaces, respectively. (<b>A</b>,<b>D</b>) Entire area. (<b>B</b>,<b>C</b>) Selected areas in (A). (<b>E</b>,<b>F</b>) Selected areas in (D).</p> Full article ">Figure 8
<p>The annual change of impervious surface maps in Daqiao New Town. Considering both the impervious-to-pervious and pervious-to-impervious changes were considered, and there may be more than one pervious-to-impervious change that occurred in a single pixel, <a href="#remotesensing-13-02409-f008" class="html-fig">Figure 8</a> shows the year of the latest pervious-to-impervious change if more than one pervious-to-impervious change was detected. In addition, <a href="#remotesensing-13-02409-f008" class="html-fig">Figure 8</a> only shows the change years for pixels, which are labeled as impervious surfaces in the year 2020. The white color represents impervious surfaces in 2006 and before; the gradient colors from shallow to deep represent the annual change of impervious surfaces.</p> Full article ">Figure 9
<p>The annual change of impervious surface maps in typical regions of Daqiao New Town shown in more detail than in <a href="#remotesensing-13-02409-f008" class="html-fig">Figure 8</a>. (<b>A</b>): Huangjia Lake University Town; (<b>B</b>): Dahualing University Town; (<b>C</b>): the athletes’ village of the 7th Military World Games, (<b>D</b>): the manufacturing industrial park. GE: Google Earth.</p> Full article ">Figure 10
<p>Accuracy of the traditional TCC model that only considers the unidirectional pervious-to-impervious conversion with the proposed STCISM that considered the bidirectional conversion.</p> Full article ">Figure 11
<p>An example showing the difference between TCC (using unidirectional pervious-to-impervious surface conversion rule) and STCISM (using the bidirectional conversion rule between pervious and impervious surface) in mapping the impervious-to-pervious surface conversion. Red and white represent impervious and pervious surfaces, respectively.</p> Full article ">
17 pages, 1947 KiB  
Review
Estimation of Forest LAI Using Discrete Airborne LiDAR: A Review
by Luo Tian, Yonghua Qu and Jianbo Qi
Remote Sens. 2021, 13(12), 2408; https://doi.org/10.3390/rs13122408 - 19 Jun 2021
Cited by 27 | Viewed by 5831
Abstract
The leaf area index (LAI) is an essential input parameter for quantitatively studying the energy and mass balance in soil-vegetation-atmosphere transfer systems. As an active remote sensing technology, light detection and ranging (LiDAR) provides a new method to describe forest canopy LAI. This [...] Read more.
The leaf area index (LAI) is an essential input parameter for quantitatively studying the energy and mass balance in soil-vegetation-atmosphere transfer systems. As an active remote sensing technology, light detection and ranging (LiDAR) provides a new method to describe forest canopy LAI. This paper reviewed the primary LAI retrieval methods using point cloud data (PCD) obtained by discrete airborne LiDAR scanner (DALS), its validation scheme, and its limitations. There are two types of LAI retrieval methods based on DALS PCD, i.e., the empirical regression and the gap fraction (GF) model. In the empirical model, tree height-related variables, LiDAR penetration indexes (LPIs), and canopy cover are the most widely used proxy variables. The height-related proxies are used most frequently; however, the LPIs proved the most efficient proxy. The GF model based on the Beer-Lambert law has been proven useful to estimate LAI; however, the suitability of LPIs is site-, tree species-, and LiDAR system-dependent. In the local validation in previous studies, poor scalability of both empirical and GF models in time, space, and across different DALS systems was observed, which means that field measurements are still needed to calibrate both types of models. The method to correct the impact from the clumping effect and woody material using DALS PCD and the saturation effect for both empirical and GF models still needs further exploration. Of most importance, further work is desired to emphasize assessing the transferability of published methods to new geographic contexts, different DALS sensors, and survey characteristics, based on figuring out the influence of each factor on the LAI retrieval process using DALS PCD. In addition, from a methodological perspective, taking advantage of DALS PCD in characterizing the 3D structure of the canopy, making full use of the ability of machine learning methods in the fusion of multisource data, developing a spatiotemporal scalable model of canopy structure parameters including LAI, and using multisource and heterogeneous data are promising areas of research. Full article
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Conceptual diagram of DALS (<b>a</b>) and preprocessing of its PCD data (<b>b</b>).</p> Full article ">Figure 2
<p>General pattern of retrieving LAI based on the gap fraction model.</p> Full article ">Figure 3
<p>Effect of <span class="html-italic">μ</span> on LPIs. Each circle marks the position of the largest difference between LPIs and <math display="inline"><semantics> <mrow> <mi>L</mi> <mi>P</mi> <msub> <mi>I</mi> <mrow> <mi>a</mi> <mi>d</mi> <mi>j</mi> <mo>.</mo> </mrow> </msub> </mrow> </semantics></math> under different <span class="html-italic">μ</span> values. The size of the circle represents the absolute difference between LPIs and <math display="inline"><semantics> <mrow> <mi>L</mi> <mi>P</mi> <msub> <mi>I</mi> <mrow> <mi>a</mi> <mi>d</mi> <mi>j</mi> <mo>.</mo> </mrow> </msub> </mrow> </semantics></math>.</p> Full article ">
22 pages, 8979 KiB  
Article
The Extraction of Street Curbs from Mobile Laser Scanning Data in Urban Areas
by Leyang Zhao, Li Yan and Xiaolin Meng
Remote Sens. 2021, 13(12), 2407; https://doi.org/10.3390/rs13122407 - 19 Jun 2021
Cited by 10 | Viewed by 3271
Abstract
The demand for mobile laser scanning in urban areas has grown in recent years. Mobile-based light detection and ranging (LiDAR) technology can be used to collect high-precision digital information on city roads and building façades. However, due to the small size of curbs, [...] Read more.
The demand for mobile laser scanning in urban areas has grown in recent years. Mobile-based light detection and ranging (LiDAR) technology can be used to collect high-precision digital information on city roads and building façades. However, due to the small size of curbs, the information that can be used for curb detection is limited. Moreover, occlusion may cause the extraction method unable to correctly capture the curb area. This paper presents the development of an algorithm for extracting street curbs from mobile-based LiDAR point cloud data to support city managers in street deformation monitoring and urban street reconstruction. The proposed method extracts curbs in three complex scenarios: vegetation covering the curbs, curved street curbs, and occlusion curbs by vehicles, pedestrians. This paper combined both spatial information and geometric information, using the spatial attributes of the road boundary. It can adapt to different heights and different road boundary structures. Analyses of real study sites show the rationality and applicability of this method for obtaining accurate results in curb-based street extraction from mobile-based LiDAR data. The overall performance of road curb extraction is fully discussed, and the results are shown to be promising. Both the completeness and correctness of the extracted left and right road edges are greater than 98%. Full article
(This article belongs to the Topic High-Resolution Earth Observation Systems, Technologies, and Applications)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>The procedure of the proposed method.</p> Full article ">Figure 2
<p>The classification algorithm of the scene elements based on their 3D geometrical properties across multiple scales.</p> Full article ">Figure 3
<p>Multiscale dimensionality classification [<a href="#B28-remotesensing-13-02407" class="html-bibr">28</a>].</p> Full article ">Figure 4
<p>The workflow of the extraction algorithm.</p> Full article ">Figure 5
<p>Elevation filtering in an urban area (<b>a</b>,<b>b</b>).</p> Full article ">Figure 6
<p>Road information relations between the sidewalk and the motorway.</p> Full article ">Figure 7
<p>Decision tree for preliminary geometric classification [<a href="#B33-remotesensing-13-02407" class="html-bibr">33</a>].</p> Full article ">Figure 8
<p>Schematic diagram of curb refinement on unconnected boundaries. (<b>a</b>) The curb which connected by a straight area. (<b>b</b>) The curb which connected by a curve fitting area.</p> Full article ">Figure 9
<p>KNN classification. (a), (b) and (c) means that the classification results of KNN are different under different K values.</p> Full article ">Figure 10
<p>The difference between KNN and RBNN in cluster ability. Subfigure (<b>a</b>) shows the data set with 2 clusters and 1 outlier, (<b>b</b>) shows 1-NN graph yielding 6 clusters, (<b>c</b>) shows 2-NN graph yielding 1 cluster and subfigure (<b>d</b>) displays RBNN graph yielding 2 clusters and 1 outlier.</p> Full article ">Figure 11
<p>The general area of the study sites.</p> Full article ">Figure 12
<p>A visual example of study sites in three complex scenes.</p> Full article ">Figure 13
<p>The visual results of multiscale dimensionality criterion classification: (<b>a</b>) Front-view images before classification; (<b>b</b>) Front-view images after classification; (<b>c</b>) Oblique view images before classification; (<b>d</b>) Oblique view images after classification; (<b>e</b>) Enlarged display of the classification results in (<b>a</b>); (<b>f</b>) Enlarged display of the classification results in (<b>b</b>); (<b>g</b>) Enlarged display of the classification results in (<b>c</b>); (<b>h</b>) Enlarged display of the classification results in (<b>d</b>).</p> Full article ">Figure 14
<p>Extraction results of the vegetation covering the curbs. (<b>a</b>,<b>b</b>) Two detail views of study sites; (<b>c</b>) Original point cloud; (<b>d</b>) Street-level imagery of the studied area overlapped by segmented curbs; (<b>e</b>) The result of curb extraction by proposed method.</p> Full article ">Figure 15
<p>Extraction results of curved curbs. (<b>a</b>,<b>b</b>) Two detail views of study sites; (<b>c</b>–<b>e</b>) Point cloud of the studied area overlapped by segmented curbs; (<b>f</b>) The result of curb extraction by proposed method.</p> Full article ">Figure 16
<p>Extraction results of occlusion curbs. (<b>a</b>,<b>e</b>) Two detail views of study sites; (<b>b</b>,<b>f</b>) The result of curb extraction by proposed method; (<b>c</b>,<b>g</b>) Original point cloud; (<b>d</b>,<b>h</b>) Point cloud of the studied area overlapped by segmented curbs.</p> Full article ">Figure 17
<p>The extraction results compared with other methods. (<b>a</b>,<b>e</b>) Proposed method in our study sites; (<b>b</b>) Yang’s extraction results in our study sites; (<b>c</b>) Kumar’s extraction results in our study sites; (<b>d</b>) Zhang’s extraction results in our study sites; (<b>f</b>) Sun’s extraction results in our study sites.</p> Full article ">
17 pages, 7701 KiB  
Article
Object-Oriented Building Contour Optimization Methodology for Image Classification Results via Generalized Gradient Vector Flow Snake Model
by Jingxin Chang, Xianjun Gao, Yuanwei Yang and Nan Wang
Remote Sens. 2021, 13(12), 2406; https://doi.org/10.3390/rs13122406 - 19 Jun 2021
Cited by 9 | Viewed by 3064
Abstract
Building boundary optimization is an essential post-process step for building extraction (by image classification). However, current boundary optimization methods through smoothing or line fitting principles are unable to optimize complex buildings. In response to this limitation, this paper proposes an object-oriented building contour [...] Read more.
Building boundary optimization is an essential post-process step for building extraction (by image classification). However, current boundary optimization methods through smoothing or line fitting principles are unable to optimize complex buildings. In response to this limitation, this paper proposes an object-oriented building contour optimization method via an improved generalized gradient vector flow (GGVF) snake model and based on the initial building contour results obtained by a classification method. First, to reduce interference from the adjacent non-building object, each building object is clipped via their extended minimum bounding rectangles (MBR). Second, an adaptive threshold Canny edge detection is applied to each building image to detect the edges, and the progressive probabilistic Hough transform (PPHT) is applied to the edge result to extract the line segments. For those cases with missing or wrong line segments in some edges, a hierarchical line segments reconstruction method is designed to obtain complete contour constraint segments. Third, accurate contour constraint segments for the GGVF snake model are designed to quickly find the target contour. With the help of the initial contour and constraint edge map for GGVF, a GGVF force field computation is executed, and the related optimization principle can be applied to complex buildings. Experimental results validate the robustness and effectiveness of the proposed method, whose contour optimization has higher accuracy and comprehensive value compared with that of the reference methods. This method can be used for effective post-processing to strengthen the accuracy of building extraction results. Full article
(This article belongs to the Special Issue A Review of Computer Vision for Remote Sensing Imagery)
►▼ Show Figures

Figure 1

Figure 1
<p>Flowchart of the building contour optimization.</p> Full article ">Figure 2
<p>Schematic diagram of obtaining the clipped building object image. (<b>a</b>) Original image; (<b>b</b>) initial extraction results (white) and minimum boundary rectangle (yellow); (<b>c</b>) expanded 3-pixel circumscribed rectangle (purple); and (<b>d</b>) cropped result after 3-pixel expansion.</p> Full article ">Figure 3
<p>Line segments from PPHT (blue) merged into one or more collinear line segments (red).</p> Full article ">Figure 4
<p>An example with intermediate results. (<b>a</b>) Original image; (<b>b</b>) initial building results; (<b>c</b>) cropped original image; (<b>d</b>) Canny detection; (<b>e</b>) PPHT segments (black line segments) and GGVF initial contour (blue closed line); and (<b>f</b>) final result.</p> Full article ">Figure 5
<p>Origin test image patches. (<b>a</b>) #1 Park City, Chicago; (<b>b</b>) #2 Hampton, Virginia; (<b>c</b>) #3 Wichita, Kansas; and (<b>d</b>) #4 Schomburg, Illinois.</p> Full article ">Figure 6
<p>Comparison of the optimization results by using the proposed method for different building extraction methods (#1-#4 is shown in <a href="#remotesensing-13-02406-f005" class="html-fig">Figure 5</a>). (<b>a</b>,<b>e</b>,<b>i</b>,<b>m</b>): building results obtained by the SSDA method; (<b>b</b>,<b>f</b>,<b>j</b>,<b>n</b>): optimization results based on SSDA; (<b>c</b>,<b>g</b>,<b>k</b>,<b>o</b>): building results obtained by the TD-GGVF method; and (<b>d</b>,<b>h</b>,<b>l</b>,<b>p</b>): optimization results based on TD-GGVF.</p> Full article ">Figure 7
<p>Comparison of different optimization methods (#1-#4 is shown in <a href="#remotesensing-13-02406-f005" class="html-fig">Figure 5</a>). (<b>a</b>,<b>e</b>,<b>i</b>,<b>m</b>): initial building results; (<b>b</b>,<b>f</b>,<b>j</b>,<b>n</b>): DPSLA method; (<b>c</b>,<b>g</b>,<b>k</b>,<b>o</b>): MCSR method; (<b>d</b>,<b>h</b>,<b>l</b>,<b>p</b>): proposed method.</p> Full article ">Figure 8
<p>Comparison of the adaptive and manual global threshold Canny edge detection results. (<b>a</b>,<b>d</b>) present the results for the original image; (<b>b</b>,<b>e</b>) present the results of the Canny edge detection with the manual global threshold; and (<b>c</b>,<b>f</b>) present the results of the Canny edge detection with the adaptive threshold.</p> Full article ">Figure 9
<p>Comparison results of the building contour constraint line segment. (<b>a</b>) Original images; (<b>b</b>) Canny edge detection; (<b>c</b>) PPHT line segments; and (<b>d</b>) optimized contour constraint line segment.</p> Full article ">Figure 10
<p>(<b>a</b>–<b>h</b>): building contour footprints; (<b>i</b>–<b>p</b>): building extraction results obtained by using the original PPHT line segments; and (<b>q</b>–<b>x</b>): building extraction results obtained by using the optimized PPHT line segments.</p> Full article ">Figure 11
<p>Accuracy comparison. (<b>a</b>) Adaptive threshold vs. manual global threshold; (<b>b</b>) optimized vs. unoptimized PPHT; (<b>c</b>) optimization proposed by this method vs. unoptimization proposed by this method; and (<b>d</b>) proposed method vs. other optimization methods.</p> Full article ">
22 pages, 2180 KiB  
Article
Enhanced Neural Network Model for Worldwide Estimation of Weighted Mean Temperature
by Fengyang Long, Chengfa Gao, Yuxiang Yan and Jinling Wang
Remote Sens. 2021, 13(12), 2405; https://doi.org/10.3390/rs13122405 - 19 Jun 2021
Cited by 5 | Viewed by 2298
Abstract
Precise modeling of weighted mean temperature (Tm) is critical for realizing real-time conversion from zenith wet delay (ZWD) to precipitation water vapor (PWV) in Global Navigation Satellite System (GNSS) meteorology applications. The empirical Tm models developed by neural network [...] Read more.
Precise modeling of weighted mean temperature (Tm) is critical for realizing real-time conversion from zenith wet delay (ZWD) to precipitation water vapor (PWV) in Global Navigation Satellite System (GNSS) meteorology applications. The empirical Tm models developed by neural network techniques have been proved to have better performances on the global scale; they also have fewer model parameters and are thus easy to operate. This paper aims to further deepen the research of Tm modeling with the neural network, and expand the application scope of Tm models and provide global users with more solutions for the real-time acquisition of Tm. An enhanced neural network Tm model (ENNTm) has been developed with the radiosonde data distributed globally. Compared with other empirical models, the ENNTm has some advanced features in both model design and model performance, Firstly, the data for modeling cover the whole troposphere rather than just near the Earth’s surface; secondly, the ensemble learning was employed to weaken the impact of sample disturbance on model performance and elaborate data preprocessing, including up-sampling and down-sampling, which was adopted to achieve better model performance on the global scale; furthermore, the ENNTm was designed to meet the requirements of three different application conditions by providing three sets of model parameters, i.e., Tm estimating without measured meteorological elements, Tm estimating with only measured temperature and Tm estimating with both measured temperature and water vapor pressure. The validation work is carried out by using the radiosonde data of global distribution, and results show that the ENNTm has better performance compared with other competing models from different perspectives under the same application conditions, the proposed model expanded the application scope of Tm estimation and provided the global users with more choices in the applications of real-time GNSS-PWV retrival. Full article
(This article belongs to the Special Issue High-Precision GNSS: Methods, Open Problems and Geoscience Applications—Part II)
►▼ Show Figures

Figure 1

Figure 1
<p>The distribution of IGRA stations used in this work.</p> Full article ">Figure 2
<p>The model structures of three different modeling schemes. The red solid circles denote the direct input parameter of BPNN models.</p> Full article ">Figure 3
<p>Box-whisker plot of MD for different modeling schemes with different BPNN structures. The small circle denotes the results after combination, the red ‘+’ is the outlier.</p> Full article ">Figure 4
<p>The MAD of different modeling schemes with different BPNN structures.</p> Full article ">Figure 5
<p>The RMSE of different modeling schemes with different BPNN structures.</p> Full article ">Figure 6
<p>The STD of different modeling schemes with different BPNN structures.</p> Full article ">Figure 7
<p>The Pearson correlation coefficient (R) of different modeling schemes with different BPNN structures.</p> Full article ">Figure 8
<p>The scattergram of MD (mean deviation) for (<b>a</b>) GTrop-Tm model; (<b>b</b>) ENNTm-A model; (<b>c</b>) NN-II model; (<b>d</b>) ENNTm-B model and (<b>e</b>) ENNTm-C model at each test IGRA station; and (<b>f</b>) their statistics in different geographical zones validated by radiosonde data during 2017 and 2018.</p> Full article ">Figure 9
<p>The scattergram of RMSE (root-mean-square error) for (<b>a</b>) GTrop-Tm model, (<b>b</b>) ENNTm-A model, (<b>c</b>) NN-II model, (<b>d</b>) ENNTm-B model and (<b>e</b>) ENNTm-C model at each test IGRA station, and (<b>f</b>) their statistics in different geographical zones validated by radiosonde data during 2017 and 2018.</p> Full article ">Figure 10
<p>The statistics of RMSE reductions in (<b>a</b>) ENNTm-A w.r.t. GTrop-Tm; (<b>b</b>) ENNTm-B w.r.t. NN-II; (<b>c</b>) ENNTm-C w.r.t. NN-II; and (<b>d</b>) ENNTm-C w.r.t. ENNTm-B. w.r.t = with respect to.</p> Full article ">Figure 11
<p>The statistics of MDs at different height layers in the (<b>a</b>) NFZ; (<b>b</b>) NTZ; (<b>c</b>) TZ; (<b>d</b>) STZ; and (<b>e</b>) SFZ.</p> Full article ">Figure 12
<p>The statistics of RMSEs at different height layers in the (<b>a</b>) NFZ; (<b>b</b>) NTZ; (<b>c</b>) TZ; (<b>d</b>) STZ; and (<b>e</b>) SFZ.</p> Full article ">Figure 13
<p>The statistics of MD for each model in different months in the (<b>a</b>) NFZ; (<b>b</b>) NTZ; (<b>c</b>) TZ; (<b>d</b>) STZ; and (<b>e</b>) SFZ.</p> Full article ">Figure 14
<p>The statistics of RMSE for each model in different months in the (<b>a</b>) NFZ; (<b>b</b>) NTZ; (<b>c</b>) TZ; (<b>d</b>) STZ; and (<b>e</b>) SFZ.</p> Full article ">
24 pages, 3935 KiB  
Article
Crop Nitrogen Retrieval Methods for Simulated Sentinel-2 Data Using In-Field Spectrometer Data
by Gregor Perich, Helge Aasen, Jochem Verrelst, Francesco Argento, Achim Walter and Frank Liebisch
Remote Sens. 2021, 13(12), 2404; https://doi.org/10.3390/rs13122404 - 19 Jun 2021
Cited by 12 | Viewed by 4731
Abstract
Nitrogen (N) is one of the key nutrients supplied in agricultural production worldwide. Over-fertilization can have negative influences on the field and the regional level (e.g., agro-ecosystems). Remote sensing of the plant N of field crops presents a valuable tool for the monitoring [...] Read more.
Nitrogen (N) is one of the key nutrients supplied in agricultural production worldwide. Over-fertilization can have negative influences on the field and the regional level (e.g., agro-ecosystems). Remote sensing of the plant N of field crops presents a valuable tool for the monitoring of N flows in agro-ecosystems. Available data for validation of satellite-based remote sensing of N is scarce. Therefore, in this study, field spectrometer measurements were used to simulate data of the Sentinel-2 (S2) satellites developed for vegetation monitoring by the ESA. The prediction performance of normalized ratio indices (NRIs), random forest regression (RFR) and Gaussian processes regression (GPR) for plant-N-related traits was assessed on a diverse real-world dataset including multiple crops, field sites and years. The plant N traits included the mass-based N measure, N concentration in the biomass (Nconc), and an area-based N measure approximating the plant N uptake (NUP). Spectral indices such as normalized ratio indices (NRIs) performed well, but the RFR and GPR methods outperformed the NRIs. Key spectral bands for each trait were identified using the RFR variable importance measure and the Gaussian processes regression band analysis tool (GPR-BAT), highlighting the importance of the short-wave infrared (SWIR) region for estimation of plant Nconc—and to a lesser extent the NUP. The red edge (RE) region was also important. The GPR-BAT showed that five bands were sufficient for plant N trait and leaf area index (LAI) estimation and that a surplus of bands effectively reduced prediction performance. A global sensitivity analysis (GSA) was performed on all traits simultaneously, showing the dominance of the LAI in the mixed remote sensing signal. To delineate the plant-N-related traits from this signal, regional and/or national data collection campaigns producing large crop spectral libraries (CSL) are needed. An improved database will likely enable the mapping of N at the agro-ecosystem level or for use in precision farming by farmers in the future. Full article
(This article belongs to the Special Issue Precision Agriculture Using Hyperspectral Images)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Coefficients of determination (R<sup>2</sup>) values for the field spectrometer (FS) dataset (empty bars) and the Sentinel-2 (S2) resampled dataset (hatched bars) for the used methods: Normalized Ratio Index (NRI, blue), Random Forest Regression (RFR, red) and Gaussian Processes Regression (GPR, green) as related to the plant traits described in <a href="#remotesensing-13-02404-t001" class="html-table">Table 1</a>.</p> Full article ">Figure 2
<p>Calculated variable importance scores of the random forest regression (RFR) on the full dataset for the field spectrometer (FS, <b>left</b>) and the Sentinel-2 (S2, <b>right</b>) resampled data. The colors show the waveband regions visible (VIS: 400–690 nm), red edge (RE: 700–790 nm), near infrared (NIR: 800–1350 nm) and short-wave infrared (SWIR: 1450–2400 nm). The water absorption bands in the regions 1350–1450, 1790–1990 and &gt;2400 nm were omitted due to their low signal to noise ratio.</p> Full article ">Figure 3
<p>Prediction performance in (R<sup>2</sup>) for the traits as a function of the number of spectral bands obtained using the Gaussian processes regression–band analysis tool (GPR-BAT) tool with sequential backward band removal (SBBR) algorithm applied (for details on SBBR, see: [<a href="#B86-remotesensing-13-02404" class="html-bibr">86</a>]).</p> Full article ">Figure 4
<p>Occurrence of the top five ranked bands with lowest GPR sigma values for the ASD sensor (<b>left</b>) and the S2 resampled sensor (<b>right</b>). Data from 10-fold cross validation, e.g., 50 (10 folds × 5 ranks) is the maximum possible occurrence.</p> Full article ">Figure 5
<p>GSA results for the ASD ground spectrometer (<b>left</b>) and the Sentinel-2 resampled (<b>right</b>) sensor for the full dataset.</p> Full article ">
16 pages, 7489 KiB  
Article
Evaluation of the J-OFURO3 Sea Surface Net Radiation and Inconsistency Correction
by Hongkai Chen, Bo Jiang, Xiuxia Li, Jianghai Peng, Hui Liang and Shaopeng Li
Remote Sens. 2021, 13(12), 2403; https://doi.org/10.3390/rs13122403 - 19 Jun 2021
Cited by 3 | Viewed by 1997
Abstract
A new satellite-based product containing daily sea surface net radiation (Rn) values at a spatial resolution of 0.25° from 1988 to 2013, named the Japanese Ocean Flux Data Sets with Use of Remote Sensing Observations, version 3 (J-OFURO3), was recently [...] Read more.
A new satellite-based product containing daily sea surface net radiation (Rn) values at a spatial resolution of 0.25° from 1988 to 2013, named the Japanese Ocean Flux Data Sets with Use of Remote Sensing Observations, version 3 (J-OFURO3), was recently generated and released. In this letter, the performance of the J-OFURO3 sea-surface Rn product was fully evaluated by using observations from 55 global moored buoy sites. The overall accuracy was satisfactory, with root-mean-square difference (RMSD) of 24.05 and 10.76 Wm−2 at daily and monthly scales, respectively. However, an inconsistency issue was found in the long-term variations in the J-OFURO3 sea-surface Rn values in approximately 2000; this inconsistency may be due to the replacement of the input dataset. To address this issue, a simple but effective inconsistency correction method was developed and conducted in this study. The analysis results demonstrated that the variations in the corrected J-OFURO3 sea-surface Rn data were more reasonable and that its daily validation accuracy was significantly improved by decreasing the bias from 4.67 to 0.27 Wm−2 before the year 2000. Thereby, it is suggested that the inconsistency correction method should be applied before using the J-OFURO3 sea-surface Rn data. However, the data users still should be cautious about another discontinuity issues caused by the quality of the input dataset itself. Full article
(This article belongs to the Special Issue Advances on Land–Ocean Heat Fluxes Using Remote Sensing)
►▼ Show Figures

Figure 1

Figure 1
<p>Distribution of 55 moored buoy sites in five measuring networks.</p> Full article ">Figure 2
<p>Flowchart of the inconsistency correction undertaken for the J-OFURO3 sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">ns</mi> </mrow> </msub> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">li</mi> </mrow> </msub> </mrow> </semantics></math> values.</p> Full article ">Figure 3
<p>Multiannual mean J-OFURO3 sea-surface <span class="html-italic">R<sub>n</sub></span> values during the period from 1988 to 2013.</p> Full article ">Figure 4
<p>Scatter plots between the J-OFURO3 sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> data and the in situ measurements at (<b>a</b>) daily and (<b>b</b>) monthly scales.</p> Full article ">Figure 5
<p>Spatial distribution of the RMSD of the J-OFURO3 sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> data at individual site at (<b>a</b>) daily and (<b>b</b>) monthly scales.</p> Full article ">Figure 5 Cont.
<p>Spatial distribution of the RMSD of the J-OFURO3 sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> data at individual site at (<b>a</b>) daily and (<b>b</b>) monthly scales.</p> Full article ">Figure 6
<p>Interannual variations in the annual mean anomalies in the sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> values recorded from 1988 to 2013 in JRA55, ERA5, MERRA2, and J-OFURO3.</p> Full article ">Figure 7
<p>The differences between the average annual mean sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> values in 1998/1999 and in 2000/2001 for the (<b>a</b>) J-OFURO3, (<b>b</b>) ISCCP-FD, (<b>c</b>) ERA5, and (<b>d</b>) NOCS 2.0 datasets.</p> Full article ">Figure 8
<p>Interannual variations in the annual mean sea-surface (<b>a</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">ns</mi> </mrow> </msub> </mrow> </semantics></math>, (<b>b</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">li</mi> </mrow> </msub> </mrow> </semantics></math>, and (<b>c</b>) <math display="inline"><semantics> <mrow> <mrow> <mi>SST</mi> <mo> </mo> </mrow> </mrow> </semantics></math>values from 1988 to 2013 in the J-OFURO3, ISCCP-FD, and CERES-3A datasets over the global ice-free oceans.</p> Full article ">Figure 9
<p>The determined pixels in the J-OFURO3 that needed to be corrected.</p> Full article ">Figure 10
<p>Interannual variations in the annual mean sea-surface (<b>a</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">ns</mi> </mrow> </msub> </mrow> </semantics></math>, (<b>b</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">li</mi> </mrow> </msub> </mrow> </semantics></math>, and (<b>c</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> <mo> </mo> </mrow> </semantics></math> from 1988 to 2013, from the original and corrected J-OFURO3 over global ice-free oceans.</p> Full article ">Figure 10 Cont.
<p>Interannual variations in the annual mean sea-surface (<b>a</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">ns</mi> </mrow> </msub> </mrow> </semantics></math>, (<b>b</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">li</mi> </mrow> </msub> </mrow> </semantics></math>, and (<b>c</b>) <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> <mo> </mo> </mrow> </semantics></math> from 1988 to 2013, from the original and corrected J-OFURO3 over global ice-free oceans.</p> Full article ">Figure 11
<p>Interannual variations in the average annual mean sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi mathvariant="italic">ns</mi> </mrow> </msub> </mrow> </semantics></math>, and <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mrow> <mi>l</mi> <mi>i</mi> </mrow> </msub> <mo> </mo> </mrow> </semantics></math> for three regions: (<b>a</b>) 20° S–20° N, (<b>b</b>) north of 20° N, and (<b>c</b>) south of 20° S from 1988 to 2013 for J-OFURO3 and corrected J-OFURO3, respectively.</p> Full article ">Figure 12
<p>Scatter plots between the J-OFURO3 daily sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> and the in situ <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math>, from 1988 to 2000, before (<b>a</b>) and after (<b>b</b>) inconsistency correction.</p> Full article ">Figure 13
<p>Times series variations in the monthly sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> in 1992 from the original corrected J-OFURO3 data and the in situ <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> at the UOP-SUB-SE moored buoy site (18° N, 22° W).</p> Full article ">Figure 14
<p>Spatiotemporal variations in the annual mean sea-surface <math display="inline"><semantics> <mrow> <msub> <mi>R</mi> <mi>n</mi> </msub> </mrow> </semantics></math> during 1988 to 2013 by linear regression model for (<b>a</b>) the corrected J-OFURO3, (<b>b</b>) the original J-OFURO3, and (<b>c</b>) the ERA5. The pixels marked with a dot indicate their statistical significance, with the <span class="html-italic">p</span>-values &lt;= 0.1.4.</p> Full article ">
15 pages, 3866 KiB  
Technical Note
New Gridded Product for the Total Columnar Atmospheric Water Vapor over Ocean Surface Constructed from Microwave Radiometer Satellite Data
by Weifu Sun, Jin Wang, Yuheng Li, Junmin Meng, Yujia Zhao and Peiqiang Wu
Remote Sens. 2021, 13(12), 2402; https://doi.org/10.3390/rs13122402 - 19 Jun 2021
Cited by 6 | Viewed by 2050
Abstract
Based on the optimal interpolation (OI) algorithm, a daily fusion product of high-resolution global ocean columnar atmospheric water vapor with a resolution of 0.25° was generated in this study from multisource remote sensing observations. The product covers the period from 2003 to 2018, [...] Read more.
Based on the optimal interpolation (OI) algorithm, a daily fusion product of high-resolution global ocean columnar atmospheric water vapor with a resolution of 0.25° was generated in this study from multisource remote sensing observations. The product covers the period from 2003 to 2018, and the data represent a fusion of microwave radiometer observations, including those from the Special Sensor Microwave Imager Sounder (SSMIS), WindSat, Advanced Microwave Scanning Radiometer for Earth Observing System sensor (AMSR-E), Advanced Microwave Scanning Radiometer 2 (AMSR2), and HY-2A microwave radiometer (MR). The accuracy of this water vapor fusion product was validated using radiosonde water vapor observations. The comparative results show that the overall mean deviation (Bias) is smaller than 0.6 mm; the root mean square error (RMSE) and standard deviation (SD) are better than 3 mm, and the mean absolute deviation (MAD) and correlation coefficient (R) are better than 2 mm and 0.98, respectively. Full article
(This article belongs to the Special Issue Advances on Land–Ocean Heat Fluxes Using Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Technical process of the water vapor fusion product.</p> Full article ">Figure 2
<p>Locations of the radiosonde stations. The red dots indicate the positions of the radiosondes.</p> Full article ">Figure 3
<p>Implementation mechanism of the fusion algorithm for water vapor. A new 0.25° background field (<span class="html-italic">B<sub>k</sub></span>) was reconstructed using the bilinear interpolation method. Then, (<span class="html-italic">O<sub>i</sub></span> − <span class="html-italic">B<sub>i</sub></span>) was calculated using the radiometer water vapor and new background.</p> Full article ">Figure 4
<p>Comparison of water vapor fusion data and single spaceborne radiometer data on 1 January 2018. The white areas in the middle and low latitudes in <a href="#remotesensing-13-02402-f004" class="html-fig">Figure 4</a> are areas that cannot be observed by satellites. The observation ability of spaceborne radiometers in polar regions is poor due to the influence of sea ice.</p> Full article ">Figure 5
<p>Daily coverage of multisource spaceborne radiometer water vapor data of 2018. The data coverage of the microwave radiometer is defined as the proportion of the effective observation data grid to the global ocean grid.</p> Full article ">Figure 6
<p>Number of effective observations of multisource spaceborne radiometer water vapor data of 2018. This figure is expressed using the daily average data calculated by using the effective water vapor observation data of spaceborne radiometers in 2018.</p> Full article ">Figure 7
<p>Comparison of the water vapor fusion products with radiosonde data from 2003 to 2018. Error distributions of the fusion product in 2003–2018 are shown in the graph on the left. The right picture shows the scatter distribution of water vapor fusion data and radiosonde data.</p> Full article ">Figure 8
<p>Taylor diagram drawn using the fusion product in 2018. The red point represents the fusion product.</p> Full article ">Figure 9
<p>Global distributions of the monthly standard deviation between the fusion product and RSS product during 2016–2018. First, the monthly water vapor of each month from 2016 to 2018 was obtained by using the fused water vapor data. Then, the calculated monthly water vapor data were compared with the monthly water vapor data of the RSS. Finally, the standard deviation was obtained by using the monthly water vapor data of 36 months.</p> Full article ">Figure 10
<p>Error comparison of fusion products using different data sources.</p> Full article ">Figure 11
<p>Error distributions of the fusion product in 2016–2018.</p> Full article ">
21 pages, 3333 KiB  
Article
Forest Restoration Monitoring Protocol with a Low-Cost Remotely Piloted Aircraft: Lessons Learned from a Case Study in the Brazilian Atlantic Forest
by Rafael Walter Albuquerque, Manuel Eduardo Ferreira, Søren Ingvor Olsen, Julio Ricardo Caetano Tymus, Cintia Palheta Balieiro, Hendrik Mansur, Ciro José Ribeiro Moura, João Vitor Silva Costa, Maurício Ruiz Castello Branco and Carlos Henrique Grohmann
Remote Sens. 2021, 13(12), 2401; https://doi.org/10.3390/rs13122401 - 19 Jun 2021
Cited by 6 | Viewed by 4034
Abstract
Traditional forest restoration (FR) monitoring methods employ spreadsheets and photos taken at the ground level. Since remotely piloted aircraft (RPA) generate a panoramic high resolution and georeferenced view of the entire area of interest, this technology has high potential to improve the traditional [...] Read more.
Traditional forest restoration (FR) monitoring methods employ spreadsheets and photos taken at the ground level. Since remotely piloted aircraft (RPA) generate a panoramic high resolution and georeferenced view of the entire area of interest, this technology has high potential to improve the traditional FR monitoring methods. This study evaluates how low-cost RPA data may contribute to FR monitoring of the Brazilian Atlantic Forest by the automatic remote measurement of Tree Density, Tree Height, Vegetation Cover (area covered by trees), and Grass Infestation. The point cloud data was processed to map the Tree Density, Tree Height, and Vegetation Cover parameters. The orthomosaic was used for a Random Forest classification that considered trees and grasses as a single land cover class. The Grass Infestation parameter was mapped by the difference between this land cover class (which considered trees and grasses) and the Vegetation Cover results (obtained by the point cloud data processing). Tree Density, Vegetation Cover, and Grass Infestation parameters presented F_scores of 0.92, 0.85, and 0.64, respectively. Tree Height accuracy was indicated by the Error Percentage considering the traditional fieldwork and the RPA results. The Error Percentage was equal to 0.13 and was considered accurate because it estimated a 13% shorter height for trees that averaged 1.93 m tall. Thus, this study showed that the FR structural parameters were accurately measured by the low-cost RPA, a technology that contributes to FR monitoring. Despite accurately measuring the structural parameters, this study reinforced the challenge of measuring the Biodiversity parameter via remote sensing because the classification of tree species was not possible. After all, the Brazilian Atlantic Forest is a biodiversity hotspot, and thus different species have similar spectral responses in the visible spectrum and similar geometric forms. Therefore, until improved automatic classification methods become available for tree species, traditional fieldwork remains necessary for a complete FR monitoring diagnostic. Full article
(This article belongs to the Special Issue Remote Sensing to Support Forest Biodiversity Assessment and Sustainable Forest Management)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location of the FR study area on Miguel Pereira municipality, situated in the Brazilian Atlantic Forest biome. To see the RPA orthomosaic and the study area on a greater scale, please go to Figure 3.</p> Full article ">Figure 2
<p>Methodology workflow of this study. MPRI is the Modified Photochemical Reflectance Index [<a href="#B23-remotesensing-13-02401" class="html-bibr">23</a>] and should be used only in the absence of shadow.</p> Full article ">Figure 3
<p>Study area divided into 8 different stretches with more, less, and mixed quantities of forested areas (FR success): Non-vegetated predominance (Nv), Non-vegetated mixed with Seedlings (NvS), Trees predominance (Tr), Trees mixed with Seedlings (TrS), Seedlings predominance (S), Trees mixed with Seedlings mixed with Vegetation Remnants (TrSVR), Erosions (Er) and Model (Md). The Md stretch was the one that best represented the whole study area in general.</p> Full article ">Figure 4
<p>There are at least six different tree species in the rectangle area, but all of them are very similar, and none could be distinguished by photointerpretation.</p> Full article ">Figure 5
<p>Individual Tree Count results of the RPA study area, which provided the Tree Density result when dividing all the identified trees by the area.</p> Full article ">Figure 6
<p>The Vegetation Cover and Grass Infestation results of the RPA study area.</p> Full article ">Figure 7
<p>The Canopy Height Model (CHM) results of the RPA study area, which provided the height of the trees that were automatically identified. The zero CHM values mean grasses or non-vegetated (bare soil) areas.</p> Full article ">Figure 8
<p>FR structural parameters results of the study area in a high mapping scale. (<b>a</b>) The RPA orthomosaic. (<b>b</b>) The Individual Tree Count, which provided the Tree Density result when dividing all the identified trees by the area. (<b>c</b>) The Vegetation Cover and Grass Infestation results. (<b>d</b>) The Canopy Height Model (CHM), which provided the height of the trees that were automatically identified and where zero CHM values means grasses or bare soil areas.</p> Full article ">Figure 9
<p>FR structural parameters of the stretches with different FR success varied from the fieldwork reference value, except Vegetation Cover. A variation in Tree Density and non-variation of Vegetation Cover suggest small tree crowns in general.</p> Full article ">Figure 10
<p>FR structural parameters Correlation Matrix between different FR stretches. In this symmetric matrix, the cells of the main diagonal shows a FR structural parameter (TD is Tree Density, VC is Vegetation Cover, TH is Tree Height, and GI is Grass Infestation) and its corresponding values in the different FR stretches. The other cells show the correlation value between different FR structural parameters, where the more asterisk (“*”) symbol occurs, the more correlated two variables are.</p> Full article ">Figure A1
<p>Vegetation Cover measuring procedure on a 25 × 4 m plot, where LT is Linear Totality. In the fieldwork procedures, LT is equal to 25 m, and the field plot is defined by considering 2 m from each LT perpendicular direction, forming then a 25 × 4 m plot. L1, L2, and L3 are examples of linear measurements in LT that are covered by trees. Thus, Vegetation Cover is the sum of L1, L2, and L3 divided by LT. Source: adapted from INEA [<a href="#B11-remotesensing-13-02401" class="html-bibr">11</a>].</p> Full article ">Figure A2
<p>An experiment to quantitatively measure the influence of the measuring tape position in Vegetation Cover acquisition if trying to replicate the fieldwork procedures on RPA imagery. Vegetation Cover value from the middle longitude of the field plot, which is where the measuring tape is positioned in the fieldwork procedures, (<b>a</b>) was compared to the Vegetation Cover value when moving the measuring tape 2 m in the west direction (<b>b</b>) and the east direction (<b>c</b>).</p> Full article ">Figure A3
<p>RPA results’ Error Percentage inside field plots when considering two reference data: photointerpretation and fieldwork.</p> Full article ">Figure A4
<p>When measuring Vegetation Cover and Grass Infestation by a 25-m line shapefile, which is a simulation of the traditional fieldwork procedure, these FR structural parameters significantly varied when moving the measuring line 2 m in east and west directions. The registered Error Percentage presented mean and standard deviation equal to 11.48 ± 39.45 for Vegetation Cover and −20.3 ± 54.61 for Grass Infestation.</p> Full article ">
15 pages, 5824 KiB  
Article
Mapping and Evaluating Human Pressure Changes in the Qilian Mountains
by Quntao Duan, Lihui Luo, Wenzhi Zhao, Yanli Zhuang and Fang Liu
Remote Sens. 2021, 13(12), 2400; https://doi.org/10.3390/rs13122400 - 19 Jun 2021
Cited by 22 | Viewed by 3735
Abstract
Human activities have dramatically changed ecosystems. As an irreplaceable ecological barrier in western China, the Qilian Mountains (QLM) provide various ecosystem services for humans. To evaluate the changes in the intensity of human activities in the QLM and their impact on the ecosystem, [...] Read more.
Human activities have dramatically changed ecosystems. As an irreplaceable ecological barrier in western China, the Qilian Mountains (QLM) provide various ecosystem services for humans. To evaluate the changes in the intensity of human activities in the QLM and their impact on the ecosystem, the human footprint (HF) method was used to conduct a spatial dataset of human activity intensity. In our study, the NDVI was used to characterize the growth of vegetation, and six categories of human pressures were employed to create the HF map in the QLM for 2000–2015 at a 1-km scale. The results showed that the mean NDVI during the growing season showed a significant increasing trend over the entire QLM in the period 2000–2015, while the NDVI showed a significant declining trend of more than 70% concentrated in Qinghai. Human pressure throughout the QLM occurred at a low level during 2000–2015, being greater in the eastern region than the western region, while the Qinghai area had greater human pressure than the Gansu area. Due to the improvement in traffic facilities, tourism, overgrazing, and other illegal activities, grasslands, shrublands, forests, wetlands, and bare land were the vegetation types most affected by human activities (in decreasing order). As the core area of the QLM, the Qilian Mountains National Nature Reserve (NR) has effectively reduced the impact of human activities. However, due to the existence of many ecological historical debts caused by unreasonable management in the past, the national park established in 2017 is facing great challenges to achieve its goals. These data and results will provide reference and guidance for future protection and restoration of the QLM ecosystem. Full article
(This article belongs to the Section Environmental Remote Sensing)
►▼ Show Figures

Figure 1

Figure 1
<p>Location of the QLM.</p> Full article ">Figure 2
<p>Land use/cover in the QLM.</p> Full article ">Figure 3
<p>HF maps of the QLM from 2000 to 2015: (<b>a</b>) HF map for 2000; (<b>b</b>) HF map for 2005; (<b>c</b>) HF map for 2010; (<b>d</b>) HF map for 2015.</p> Full article ">Figure 4
<p>The proportion of different human pressure classes in 2000, 2005, 2010, and 2015.</p> Full article ">Figure 5
<p>Change in the HF over the QLM: (<b>a</b>) change in the HF for 2000–2005; (<b>b</b>) change in the HF for 2005–2015; (<b>c</b>) change in the HF for 2000–2015.</p> Full article ">Figure 6
<p>Growing season NDVI trends obtained by linear regression, and significance levels obtained by Mann–Kendall tests for the QLM from 2000 to 2015: (<b>a</b>) NDVI trend from 2000 to 2005; (<b>b</b>) NDVI trend from 2005 to 2015; (<b>c</b>) significance level for the NDVI from 2000 to 2005; (<b>d</b>) significance level for the NDVI from 2005 to 2015. The blank regions in the QLM in c and d indicate a nonsignificant change in NDVI (<span class="html-italic">p</span> &gt; 0.1).</p> Full article ">Figure 7
<p>Relationships between significant NDVI trends and changes in the HF for 2000–2005 and 2005–2015 in Gansu and Qinghai. The blue lines show the linear regression results.</p> Full article ">Figure 8
<p>Relationships between the trends of different vegetation types and changes in the HF and six categories of human pressures for 2000–2005 and 2005–2015 in Gansu and Qinghai. The blue regions indicate significant positive correlations (<span class="html-italic">p</span> &lt; 0.05), and the correlation coefficient increases with darkening color; the red regions indicate significant positive correlations (<span class="html-italic">p</span> &lt; 0.05), and the correlation coefficient increases with darkening color. Gray regions denote nonsignificant correlations (<span class="html-italic">p</span> &gt; 0.05). The number in the bin is the correlation coefficient.</p> Full article ">
19 pages, 2903 KiB  
Article
HORUS: Multispectral and Multiangle CubeSat Mission Targeting Sub-Kilometer Remote Sensing Applications
by Alice Pellegrino, Maria Giulia Pancalli, Andrea Gianfermo, Paolo Marzioli, Federico Curianò, Federica Angeletti, Fabrizio Piergentili and Fabio Santoni
Remote Sens. 2021, 13(12), 2399; https://doi.org/10.3390/rs13122399 - 19 Jun 2021
Cited by 5 | Viewed by 4494
Abstract
This paper presents the HORUS mission, aimed at multispectral and multiangle (nadir and off-nadir) planetary optical observation, using Commercial Off-The-Shelf (COTS) instruments on-board a 6-Unit CubeSat. The collected data are characterized by a sub-kilometer resolution, useful for different applications for environmental monitoring, atmospheric [...] Read more.
This paper presents the HORUS mission, aimed at multispectral and multiangle (nadir and off-nadir) planetary optical observation, using Commercial Off-The-Shelf (COTS) instruments on-board a 6-Unit CubeSat. The collected data are characterized by a sub-kilometer resolution, useful for different applications for environmental monitoring, atmospheric characterization, and ocean studies. Latest advancements in electro-optical instrumentation permit to consider an optimized instrument able to fit in a small volume, in principle without significant reduction in the achievable performances with respect to typical large-spacecraft implementations. CubeSat-based platforms ensure high flexibility, with fast and simple components’ integration, and may be used as stand-alone system or in synergy with larger missions, for example to improve revisit time. The mission rationale, its main objectives and scientific background, including the combination of off-nadir potential continuous multiangle coverage in a full perspective and related observation bands are provided. The observation system conceptual design and its installation on-board a 6U CubeSat bus, together with the spacecraft subsystems are discussed, assessing the feasibility of the mission and its suitability as a building block for a multiplatform distributed system. Full article
(This article belongs to the Special Issue CubeSats Applications for Earth and Prospectives for Planetary Remote Sensing)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Acquisition geometry of a matrix imager, with HFOV aligned along-track.</p> Full article ">Figure 2
<p>2D representation of HORUS iFOV.</p> Full article ">Figure 3
<p>The HORUS camera quadruplet (Types a-b-c-d) FOVs and off-nadir boresight angles (not to scale).</p> Full article ">Figure 4
<p>Geometrical relation between view angle and on-board off-nadir boresight angle.</p> Full article ">Figure 5
<p>Schematic representation of NASA-JPL MISR view angles, all within the view-angle range reached by HORUS (not to scale).</p> Full article ">Figure 6
<p>HORUS CAD model.</p> Full article ">Figure 7
<p>Cartesian projections of the HORUS spacecraft CAD model.</p> Full article ">
26 pages, 54684 KiB  
Article
Fast Unsupervised Multi-Scale Characterization of Urban Landscapes Based on Earth Observation Data
by Claire Teillet, Benjamin Pillot, Thibault Catry, Laurent Demagistri, Dominique Lyszczarz, Marc Lang, Pierre Couteron, Nicolas Barbier, Arsène Adou Kouassi, Quentin Gunther and Nadine Dessay
Remote Sens. 2021, 13(12), 2398; https://doi.org/10.3390/rs13122398 - 19 Jun 2021
Cited by 3 | Viewed by 2882
Abstract
Most remote sensing studies of urban areas focus on a single scale, using supervised methodologies and very few analyses focus on the “neighborhood” scale. The lack of multi-scale analysis, together with the scarcity of training and validation datasets in many countries lead us [...] Read more.
Most remote sensing studies of urban areas focus on a single scale, using supervised methodologies and very few analyses focus on the “neighborhood” scale. The lack of multi-scale analysis, together with the scarcity of training and validation datasets in many countries lead us to propose a single fast unsupervised method for the characterization of urban areas. With the FOTOTEX algorithm, this paper introduces a texture-based method to characterize urban areas at three nested scales: macro-scale (urban footprint), meso-scale (“neighbourhoods”) and micro-scale (objects). FOTOTEX combines a Fast Fourier Transform and a Principal Component Analysis to convert texture into frequency signal. Several parameters were tested over Sentinel-2 and Pleiades imagery on Bouake and Brasilia. Results showed that a single Sentinel-2 image better assesses the urban footprint than the global products. Pleiades images allowed discriminating neighbourhoods and urban objects using texture, which is correlated with metrics such as building density, built-up and vegetation proportions. The best configurations for each scale of analysis were determined and recommendations provided to users. The open FOTOTEX algorithm demonstrated a strong potential to characterize the three nested scales of urban areas, especially when training and validation data are scarce, and computing resources limited. Full article
(This article belongs to the Special Issue Remote Sensing Approaches to Landscape Analysis of Urban and Peri-Urban Environments)
►▼ Show Figures

Figure 1

Figure 1
<p>Study areas and satellite imagery used in this study. On the left, band 2 (blue) of Sentinel 2 images on Bouake (<b>top</b>) and Brasilia (<b>bottom</b>) are shown (Copernicus Sentinel data (2018-2020)/ESA). On the right Pleiades images (panchromatic channel) on the northeastern part of Bouake and Sao Sebastiao, a satellite City of Brasilia (Contains information © CNES 2020, Distribution Airbus DS, all rights reserved. No commercial use.).</p> Full article ">Figure 2
<p>Potential of FOTOTEX for the multi-scale characterization of urban landscapes, with the scales of analysis and associated Earth observation data.</p> Full article ">Figure 3
<p>Simplified methodological framework of the FOTOTEX algorithm, divided in four steps (<b>a</b>) image partitioning, (<b>b</b>) spectral analysis by Fourier transform and R-spectra computation, (<b>c</b>) Principal Component Analysis and (<b>d</b>) RGB composite of the three main components (textural indices). These steps will be described in detail in the method section below.</p> Full article ">Figure 4
<p>Influence of spectral bands used as input in the FOTOTEX algorithm for the characterization of urban areas at the macro-scale using Sentinel 2 images on Bouake. (<b>a</b>) B2 (blue) sentinel 2 band, and (<b>b</b>) B6 (NIR) Sentinel 2 band. (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 5
<p>Influence of the window size parameter in the FOTOTEX algorithm for the characterization of urban areas at the macro-scale using Sentinel 2 images on Bouake and Brasilia. (<b>a</b>) B2 (blue) sentinel 2 band on Bouake with a window size of 5 pixels, (<b>b</b>) B2 (blue) sentinel 2 band on Bouake with a window size of 31 pixels, (<b>c</b>) B2 (blue) sentinel 2 band on Brasilia with a window size of 5 pixels and (<b>d</b>) B2 (blue) sentinel 2 band on Brasilia with a window size of 31 pixels. (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 6
<p>Influence of the dc component and the normalization parameters in the FOTOTEX algorithm for the characterization of urban areas at the macro-scale using Sentinel 2 images on Bouake. (<b>a</b>) B2 (blue) sentinel 2 band on Bouake with DC component (DC = True) and without normalization (N = False), (<b>b</b>) B2 (blue) sentinel 2 band on Bouake without DC component (DC = False) and without normalization (N = False) and (<b>c</b>) B2 (blue) sentinel 2 band on Bouake with DC component (DC = True) and normalization (N = True). The text in red highlights the modified parameter compared with image (<b>a</b>). (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 7
<p>Comparison of urban footprints extracted from Sentinel 2 images on Bouake and Brasilia with FOTOTEX with the Global Human Settlement Layer (GHSL) product from JRC (© European Union) over different window sizes, (<b>a</b>) urban footprints on Bouake window size = 5 pixels, (<b>b</b>) urban footprints on Bouake window size = 31 pixels, (<b>c</b>) urban footprints on Brasilia window size = 5 pixels, (<b>d</b>) urban footprints on Brasilia window size = 31 pixels, Sentinel 2 Background © ESA (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 8
<p>Influence of various parameters (partitioning method, pixel size and window size) in FOTOTEX in the characterization of urban units at the meso-scale from Pleiades images on the northeastern part of Bouake and Sao Sebastiao, Brasilia. (<b>a</b>,<b>e</b>) use of block method, (<b>b</b>,<b>f</b>) use of moving window method, (<b>c</b>,<b>g</b>) use of a 3 m input pixel size over, (<b>d</b>,<b>h</b>) use of a window size of 31 pixels. The text in red highlights the modified parameter compared with image (<b>a</b>). (PX = pixel size, WS = window size, PM = partitioning method, MW = Moving Window, DC = DC component, N = normalize).</p> Full article ">Figure 9
<p>Influence of DC component and normalization in FOTOTEX in the characterization of urban units at the meso-scale from Pleiades images on the northeastern part of Bouake and Sao Sebastiao, Brasilia. (<b>a</b>,<b>d</b>) use of block method over Bouake and Brasilia, with DC component, without normalization, (<b>b</b>,<b>e</b>) use of block method over Bouake and Brasilia, without DC component, without normalization, (<b>c</b>,<b>f</b>) use of block method over Bouake and Brasilia, with DC component and normalization. The text in red highlights the modified parameter compared with image (<b>a</b>). (PX = pixel size, WS = window size, PM = partitioning method, MW = Moving Window, DC = DC component, N = normalize).</p> Full article ">Figure 10
<p>Study of the relationship between texture information at the meso-scale extracted using FOTOTEX on Pleiades images on Sao Sebastiao, Brasilia (<b>a</b>) and the northeastern part of Bouake (<b>c</b>), with the density of buildings represented as heatmaps using building contours from the SEDUH database (<a href="http://www.seduh.df.gov.br/" target="_blank">http://www.seduh.df.gov.br/</a> accessed on 20 September 2020) for Brasilia (<b>b</b>) and OpenStreetMap for Bouake (<b>d</b>). On this figure, black/colored lines and numbers are urban units extracted from the combination of texture and contours from the urban footprint produced with FOTOTEX and Sentinel 2.</p> Full article ">Figure 11
<p>Influence of the partitioning method for the micro-scale characterization of urban objects on Bouake and Brasilia using Pleiades images (<b>a</b>) block method on Bouake, (<b>b</b>) block method on Brasilia, (<b>c</b>) moving window method on Bouake and (<b>d</b>) moving window method on Brasilia (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 12
<p>Implementation of the FOTOTEX algorithm on (<b>a</b>) a UAV image captured over the city of Bouake (<b>b</b>) showing the RGB composite produced by FOTOTEX (<b>c</b>) and the three Principal Components. (PX = pixel size, WS = window size, PM = partitioning method, DC = DC component, N = normalize).</p> Full article ">Figure 13
<p>Extraction of buildings edges with (<b>a</b>) a segmentation method carried out with the Large Scale Generic Region Merging algorithm (Orfeo Toolbox) and (<b>b</b>) the FOTOTEX method implemented on the blue band of UAV image. (<b>c</b>,<b>d</b>) zooms on two different areas to compare detection by FOTOTEX and segmentation.</p> Full article ">
19 pages, 5213 KiB  
Article
Application of RGB Images Obtained by UAV in Coffee Farming
by Brenon Diennevam Souza Barbosa, Gabriel Araújo e Silva Ferraz, Luana Mendes dos Santos, Lucas Santos Santana, Diego Bedin Marin, Giuseppe Rossi and Leonardo Conti
Remote Sens. 2021, 13(12), 2397; https://doi.org/10.3390/rs13122397 - 19 Jun 2021
Cited by 22 | Viewed by 4505
Abstract
The objective of this study was to evaluate the potential of the practical application of unmanned aerial vehicles and RGB vegetation indices (VIs) in the monitoring of a coffee crop. The study was conducted in an experimental coffee field over a 12-month period. [...] Read more.
The objective of this study was to evaluate the potential of the practical application of unmanned aerial vehicles and RGB vegetation indices (VIs) in the monitoring of a coffee crop. The study was conducted in an experimental coffee field over a 12-month period. An RGB digital camera coupled to a UAV was used. Nine VIs were evaluated in this study. These VIs were subjected to a Pearson correlation analysis with the leaf area index (LAI), and subsequently, the VIs with higher R2 values were selected. The LAI was estimated by plant height and crown diameter values obtained by imaging, which were correlated with these values measured in the field. Among the VIs evaluated, MPRI (0.31) and GLI (0.41) presented greater correlation with LAI; however, the correlation was weak. Thematic maps of VIs in the evaluated period showed variability present in the crop. The evolution of weeds in the planting rows was noticeable with both VIs, which can help managers to make the decision to start crop management, thus saving resources. The results show that the use of low-cost UAVs and RGB cameras has potential for monitoring the coffee production cycle, providing producers with information in a more accurate, quick and simple way. Full article
(This article belongs to the Special Issue Remote Sensing for Smart Agriculture Management)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Location of the study area.</p> Full article ">Figure 2
<p>Flowchart for calculating the VI and correlation with LAI (<b>A</b>). Buffer of 0.8 m to calculate the mean value of VI in each sample (<b>B</b>).</p> Full article ">Figure 3
<p>Outlier example.</p> Full article ">Figure 4
<p>Estimated x measured plant heights in the field (<b>A</b>). Estimated x measured plant crown diameters (<b>B</b>).</p> Full article ">Figure 5
<p>Behaviour of the LAI estimated with the UAV.</p> Full article ">Figure 6
<p>Behaviour of MPRI, GLI and LAI in the coffee production cycle.</p> Full article ">Figure 7
<p>Second phase of the vegetative cycle of coffee evidenced by the VARI: (<b>A</b>)—VARI/june, (<b>B</b>)—VARI/july, (<b>C</b>)—VARI/august.</p> Full article ">Figure 8
<p>Third phase of the vegetative cycle of coffee evidenced by RGBVI: (<b>A</b>) —RGBVI/september, (<b>B</b>)—RGBVI/october, (<b>C</b>)—RGBVI/november, (<b>D</b>)—RGBVI/december.</p> Full article ">Figure 9
<p>Fourth phase of the vegetative cycle of coffee evidenced by MPRI: (<b>A</b>)—MPRI/january, (<b>B</b>)—MPRI/february, (<b>C</b>)—MPRI/march.</p> Full article ">Figure 10
<p>Fifth phase of the vegetative cycle of coffee evidenced by GLI: (<b>A</b>)—GLI/april, (<b>B</b>)—GLI/may.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-192
Previous Issue
Next Issue
Back to TopTop