Remote Sensing

Research

19 pages, 1910 KiB

Open AccessArticle

Detection of Glass Insulators Using Deep Neural Networks Based on Optical Imaging

by Jinyu Wang, Yingna Li and Wenxiang Chen

Remote Sens. 2022, 14(20), 5153; https://doi.org/10.3390/rs14205153 - 15 Oct 2022

Cited by 11 | Viewed by 2335

As the pre-part of tasks such as fault detection and line inspection, insulator detection is a crucial task. However, considering the complex environment of high-voltage transmission lines, the traditional insulator detection accuracy is unsatisfactory, and manual inspection is dangerous and inefficient. To improve [...] Read more.

As the pre-part of tasks such as fault detection and line inspection, insulator detection is a crucial task. However, considering the complex environment of high-voltage transmission lines, the traditional insulator detection accuracy is unsatisfactory, and manual inspection is dangerous and inefficient. To improve this situation, this paper proposes an insulator detection model Siamese ID-YOLO based on a deep neural network. The model achieves the best balance between speed and accuracy compared with traditional detection methods. In order to achieve the purpose of image enhancement, this paper adopts the canny-based edge detection operator to highlight the edges of insulators to obtain more semantic information. In this paper, based on the Darknet53 network and Siamese network, the insulator original image and the edge image are jointly input into the model. Siamese IN-YOLO model achieves more fine-grained extraction of insulators through weight sharing between Siamese networks, thereby improving the detection accuracy of insulators. This paper uses statistical clustering analysis on the area and aspect ratio of the insulator data set, then pre-set and adjusts the hyperparameters of the model anchor box to make it more suitable for the insulator detection task. In addition, this paper makes an insulator dataset named InsuDaSet based on UAV(Unmanned Aerial Vehicle) shoot insulator images for model training. The experiments show that the insulator detection can reach 92.72% detection accuracy and 84FPS detection speed, which can fully meet the online insulator detection requirements. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

30 pages, 11233 KiB

Open AccessArticle

Multibranch Unsupervised Domain Adaptation Network for Cross Multidomain Orchard Area Segmentation

by Ming Liu, Dong Ren, Hang Sun and Simon X. Yang

Remote Sens. 2022, 14(19), 4915; https://doi.org/10.3390/rs14194915 - 1 Oct 2022

Cited by 1 | Viewed by 1841

Abstract

Although unsupervised domain adaptation (UDA) has been extensively studied in remote sensing image segmentation tasks, most UDA models are designed based on single-target domain settings. Large-scale remote sensing images often have multiple target domains in practical applications, and the simple extension of single-target [...] Read more.

Although unsupervised domain adaptation (UDA) has been extensively studied in remote sensing image segmentation tasks, most UDA models are designed based on single-target domain settings. Large-scale remote sensing images often have multiple target domains in practical applications, and the simple extension of single-target UDA models to multiple target domains is unstable and costly. Multi-target unsupervised domain adaptation (MTUDA) is a more practical scenario that has great potential for solving the problem of crossing multiple domains in remote sensing images. However, existing MTUDA models neglect to learn and control the private features of the target domain, leading to missing information and negative migration. To solve these problems, this paper proposes a multibranch unsupervised domain adaptation network (MBUDA) for orchard area segmentation. The multibranch framework aligns multiple domain features, while preventing private features from interfering with training. We introduce multiple ancillary classifiers to help the model learn more robust latent target domain data representations. Additionally, we propose an adaptation enhanced learning strategy to reduce the distribution gaps further and enhance the adaptation effect. To evaluate the proposed method, this paper utilizes two settings with different numbers of target domains. On average, the proposed method achieves a high IoU gain of 7.47% over the baseline (single-target UDA), reducing costs and ensuring segmentation model performance in multiple target domains. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

20 pages, 11207 KiB

Open AccessArticle

Optimization of Remote Sensing Image Segmentation by a Customized Parallel Sine Cosine Algorithm Based on the Taguchi Method

by Fang Fan, Gaoyuan Liu, Jiarong Geng, Huiqi Zhao and Gang Liu

Remote Sens. 2022, 14(19), 4875; https://doi.org/10.3390/rs14194875 - 29 Sep 2022

Cited by 7 | Viewed by 2259

Abstract

Affected by solar radiation, atmospheric windows, radiation aberrations, and other air and sky environmental factors, remote sensing images usually contain a large amount of noise and suffer from problems such as non-uniform image feature density. These problems bring great difficulties to the segmentation [...] Read more.

Affected by solar radiation, atmospheric windows, radiation aberrations, and other air and sky environmental factors, remote sensing images usually contain a large amount of noise and suffer from problems such as non-uniform image feature density. These problems bring great difficulties to the segmentation of high-precision remote sensing image. To improve the segmentation effect of remote sensing images, this study adopted an improved metaheuristic algorithm to optimize the parameter settings of pulse-coupled neural networks (PCNNs). Using the Taguchi method, the optimal parallelism scheme of the algorithm was effectively tailored for a specific target problem. The blindness in the design of the algorithm parallel structure was effectively avoided. The superiority of the customized parallel SCA based on the Taguchi method (TPSCA) was demonstrated in tests with different types of benchmark functions. In this study, simulations were performed using IKONOS, GeoEye-1, and WorldView-2 satellite remote sensing images. The results showed that the accuracy of the proposed remote sensing image segmentation model was significantly improved. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Graphical abstract

23 pages, 18464 KiB

Open AccessArticle

The Self-Supervised Spectral–Spatial Vision Transformer Network for Accurate Prediction of Wheat Nitrogen Status from UAV Imagery

by Xin Zhang, Liangxiu Han, Tam Sobeih, Lewis Lappin, Mark A. Lee, Andew Howard and Aron Kisdi

Remote Sens. 2022, 14(6), 1400; https://doi.org/10.3390/rs14061400 - 14 Mar 2022

Cited by 16 | Viewed by 6680

Abstract

Nitrogen (N) fertilizer is routinely applied by farmers to increase crop yields. At present, farmers often over-apply N fertilizer in some locations or at certain times because they do not have high-resolution crop N status data. N-use efficiency can be low, with the [...] Read more.

Nitrogen (N) fertilizer is routinely applied by farmers to increase crop yields. At present, farmers often over-apply N fertilizer in some locations or at certain times because they do not have high-resolution crop N status data. N-use efficiency can be low, with the remaining N lost to the environment, resulting in higher production costs and environmental pollution. Accurate and timely estimation of N status in crops is crucial to improving cropping systems’ economic and environmental sustainability. Destructive approaches based on plant tissue analysis are time consuming and impractical over large fields. Recent advances in remote sensing and deep learning have shown promise in addressing the aforementioned challenges in a non-destructive way. In this work, we propose a novel deep learning framework: a self-supervised spectral–spatial attention-based vision transformer (SSVT). The proposed SSVT introduces a Spectral Attention Block (SAB) and a Spatial Interaction Block (SIB), which allows for simultaneous learning of both spatial and spectral features from UAV digital aerial imagery, for accurate N status prediction in wheat fields. Moreover, the proposed framework introduces local-to-global self-supervised learning to help train the model from unlabelled data. The proposed SSVT has been compared with five state-of-the-art models including: ResNet, RegNet, EfficientNet, EfficientNetV2, and the original vision transformer on both testing and independent datasets. The proposed approach achieved high accuracy (0.96) with good generalizability and reproducibility for wheat N status estimation. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

18 pages, 29045 KiB

Open AccessArticle

Integrating Remote Sensing and Meteorological Data to Predict Wheat Stripe Rust

by Chao Ruan, Yingying Dong, Wenjiang Huang, Linsheng Huang, Huichun Ye, Huiqin Ma, Anting Guo and Ruiqi Sun

Remote Sens. 2022, 14(5), 1221; https://doi.org/10.3390/rs14051221 - 2 Mar 2022

Cited by 12 | Viewed by 3824

Abstract

Wheat stripe rust poses a serious threat to wheat production. An effective prediction method is important for food security. In this study, we developed a prediction model for wheat stripe rust based on vegetation indices and meteorological features. First, based on time-series Sentinel-2 [...] Read more.

Wheat stripe rust poses a serious threat to wheat production. An effective prediction method is important for food security. In this study, we developed a prediction model for wheat stripe rust based on vegetation indices and meteorological features. First, based on time-series Sentinel-2 remote sensing images and meteorological data, wheat phenology (jointing date) was estimated using the harmonic analysis of time-series combined with average cumulative temperature. Then, vegetation indices were extracted based on phenological information. Meteorological features were screened using correlation analysis combined with independent t-test analysis. Finally, a random forest (RF) was used to construct a prediction model for wheat stripe rust. The results showed that the RF model using the input combination (phenological information-based vegetation indices and meteorological features) produced a higher prediction accuracy and a kappa coefficient of 88.7% and 0.772, respectively. The prediction model using phenological information-based vegetation indices outperformed the prediction model using single-date image-based vegetation indices, and the overall accuracy improved from 62.9% to 78.4%. These results indicated that the method combining phenological information-based vegetation indices and meteorological features can be used for wheat stripe rust prediction. The results of the prediction model can provide guidance and suggestions for disease prevention in the study area. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

21 pages, 4366 KiB

Open AccessArticle

Dynamic Forecast of Desert Locust Presence Using Machine Learning with a Multivariate Time Lag Sliding Window Technique

by Ruiqi Sun, Wenjiang Huang, Yingying Dong, Longlong Zhao, Biyao Zhang, Huiqin Ma, Yun Geng, Chao Ruan, Naichen Xing, Xidong Chen and Xueling Li

Remote Sens. 2022, 14(3), 747; https://doi.org/10.3390/rs14030747 - 5 Feb 2022

Cited by 14 | Viewed by 3853

Abstract

Desert locust plagues can easily cause a regional food crisis and thus affect social stability. Preventive control of the disaster highlights the early detection of hopper gregarization before they form devastating swarms. However, the response of hopper band emergence to environmental fluctuation exhibits [...] Read more.

Desert locust plagues can easily cause a regional food crisis and thus affect social stability. Preventive control of the disaster highlights the early detection of hopper gregarization before they form devastating swarms. However, the response of hopper band emergence to environmental fluctuation exhibits a time lag. To realize the dynamic forecast of band occurrence with optimal temporal predictors, we proposed an SVM-based model with a temporal sliding window technique by coupling multisource time-series imagery with historical locust ground survey observations from between 2000–2020. The sliding window method was based on a lagging variable importance ranking used to analyze the temporal organization of environmental indicators in band-forming sequences and eventually facilitate the early prediction of band emergence. Statistical results show that hopper bands are more likely to occur within 41–64 days after increased rainfall; soil moisture dynamics increasing by approximately 0.05 m³/m³ then decreasing may enhance the chance of observing bands after 73–80 days. While sparse vegetation areas with NDVI increasing from 0.18 to 0.25 tend to witness bands after 17–40 days. The forecast model combining the optimal time lags of these dynamic indicators with other static indicators allows for a 16-day extended outlook of band presence in Somalia, Ethiopia, and Kenya. Monthly predictions from February to December 2020 display an overall accuracy of 77.46%, with an average ROC-AUC of 0.767 and a mean F-score close to 0.772. The multivariate forecast framework based on the lagging effect can realize the early warning of band presence in different spatiotemporal scenarios, supporting early decisions and response strategies for desert locust preventive management. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

22 pages, 8707 KiB

Open AccessArticle

Novel CropdocNet Model for Automated Potato Late Blight Disease Detection from Unmanned Aerial Vehicle-Based Hyperspectral Imagery

by Yue Shi, Liangxiu Han, Anthony Kleerekoper, Sheng Chang and Tongle Hu

Remote Sens. 2022, 14(2), 396; https://doi.org/10.3390/rs14020396 - 15 Jan 2022

Cited by 46 | Viewed by 4632

Abstract

The accurate and automated diagnosis of potato late blight disease, one of the most destructive potato diseases, is critical for precision agricultural control and management. Recent advances in remote sensing and deep learning offer the opportunity to address this challenge. This study proposes [...] Read more.

The accurate and automated diagnosis of potato late blight disease, one of the most destructive potato diseases, is critical for precision agricultural control and management. Recent advances in remote sensing and deep learning offer the opportunity to address this challenge. This study proposes a novel end-to-end deep learning model (CropdocNet) for accurate and automated late blight disease diagnosis from UAV-based hyperspectral imagery. The proposed method considers the potential disease-specific reflectance radiation variance caused by the canopy’s structural diversity and introduces multiple capsule layers to model the part-to-whole relationship between spectral–spatial features and the target classes to represent the rotation invariance of the target classes in the feature space. We evaluate the proposed method with real UAV-based HSI data under controlled and natural field conditions. The effectiveness of the hierarchical features is quantitatively assessed and compared with the existing representative machine learning/deep learning methods on both testing and independent datasets. The experimental results show that the proposed model significantly improves accuracy when considering the hierarchical structure of spectral–spatial features, with average accuracies of 98.09% for the testing dataset and 95.75% for the independent dataset, respectively. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

27 pages, 24644 KiB

Open AccessArticle

Spatiotemporal Evolution Analysis and Future Scenario Prediction of Rocky Desertification in a Subtropical Karst Region

by Chunhua Qian, Hequn Qiang, Changyou Qin, Zi Wang and Mingyang Li

Remote Sens. 2022, 14(2), 292; https://doi.org/10.3390/rs14020292 - 9 Jan 2022

Cited by 12 | Viewed by 2851

Abstract

Landscape change is a dynamic feature of landscape structure and function over time which is usually affected by natural and human factors. The evolution of rocky desertification is a typical landscape change that directly affects ecological environment governance and sustainable development. Guizhou is [...] Read more.

Landscape change is a dynamic feature of landscape structure and function over time which is usually affected by natural and human factors. The evolution of rocky desertification is a typical landscape change that directly affects ecological environment governance and sustainable development. Guizhou is one of the most typical subtropical karst landform areas in the world. Its special karst rocky desertification phenomenon is an important factor affecting the ecological environment and limiting sustainable development. In this paper, remote sensing imagery and machine learning methods are utilized to model and analyze the spatiotemporal variation of rocky desertification in Guizhou. Based on an improved CA-Markov model, rocky desertification scenarios in the next 30 years are predicted, providing data support for exploration of the evolution rule of rocky desertification in subtropical karst areas and for effective management. The specific results are as follows: (1) Based on the dynamic degree, transfer matrix, evolution intensity, and speed, the temporal and spatial evolution of rocky desertification in Guizhou from 2001 to 2020 was analyzed. It was found that the proportion of no rocky desertification (NRD) areas increased from 48.86% to 63.53% over this period. Potential rocky desertification (PRD), light rocky desertification (LRD), middle rocky desertification (MRD), and severe rocky desertification (SRD) continued to improve, with the improvement showing an accelerating trend after 2010. (2) An improved CA-Markov model was used to predict the future rocky desertification scenario; compared to the traditional CA-Markov model, the Lee–Sallee index increased from 0.681 to 0.723, and figure of merit (FOM) increased from 0.459 to 0.530. The conclusions of this paper are as follows: (1) From 2001 to 2020, the evolution speed of PRD was the fastest, while that of SRD was the slowest. Rocky desertification control should not only focus on areas with serious rocky desertification, but also prevent transformation from NRD to PRD. (2) Rocky desertification will continue to improve over the next 30 years. Possible deterioration areas are concentrated in high-altitude areas, such as the south of Bijie and the east of Liupanshui. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

21 pages, 1899 KiB

Open AccessArticle

A Random Forest Algorithm for Retrieving Canopy Chlorophyll Content of Wheat and Soybean Trained with PROSAIL Simulations Using Adjusted Average Leaf Angle

by Quanjun Jiao, Qi Sun, Bing Zhang, Wenjiang Huang, Huichun Ye, Zhaoming Zhang, Xiao Zhang and Binxiang Qian

Remote Sens. 2022, 14(1), 98; https://doi.org/10.3390/rs14010098 - 25 Dec 2021

Cited by 33 | Viewed by 4487

Abstract

Canopy chlorophyll content (CCC) is an important indicator for crop-growth monitoring and crop productivity estimation. The hybrid method, involving the PROSAIL radiative transfer model and machine learning algorithms, has been widely applied for crop CCC retrieval. However, PROSAIL’s homogeneous canopy hypothesis limits the [...] Read more.

Canopy chlorophyll content (CCC) is an important indicator for crop-growth monitoring and crop productivity estimation. The hybrid method, involving the PROSAIL radiative transfer model and machine learning algorithms, has been widely applied for crop CCC retrieval. However, PROSAIL’s homogeneous canopy hypothesis limits the ability to use the PROSAIL-based CCC estimation across different crops with a row structure. In addition to leaf area index (LAI), average leaf angle (ALA) is the most important canopy structure factor in the PROSAIL model. Under the same LAI, adjustment of the ALA can make a PROSAIL simulation obtain the same canopy gap as the heterogeneous canopy at a specific observation angle. Therefore, parameterization of an adjusted ALA (ALA_adj) is an optimal choice to make the PROSAIL model suitable for specific row-planted crops. This paper attempted to improve PROSAIL-based CCC retrieval for different crops, using a random forest algorithm, by introducing the prior knowledge of crop-specific ALA_adj. Based on the field reflectance spectrum at nadir, leaf area index, and leaf chlorophyll content, parameterization of the ALA_adj in the PROSAIL model for wheat and soybean was carried out. An algorithm integrating the random forest and PROSAIL simulations with prior ALA_adj information was developed for wheat and soybean CCC retrieval. Ground-measured CCC measurements were used to validate the CCC retrieved from canopy spectra. The results showed that the ALA_adj values (62 degrees for wheat; 45 degrees for soybean) that were parameterized for the PROSAIL model demonstrated good discrimination between the two crops. The proposed algorithm improved the CCC retrieval accuracy for wheat and soybean, regardless of whether continuous visible to near-infrared spectra with 50 bands (RMSE from 39.9 to 32.9 μg cm⁻²; R² from 0.67 to 0.76) or discrete spectra with 13 bands (RMSE from 43.9 to 33.7 μg cm⁻²; R² from 0.63 to 0.74) and nine bands (RMSE from 45.1 to 37.0 μg cm⁻²; R² from 0.61 to 0.71) were used. The proposed hybrid algorithm, based on PROSAIL simulations with ALA_adj, has the potential for satellite-based CCC estimation across different crop types, and it also has a good reference value for the retrieval of other crop parameters. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
Workflow of crop canopy chlorophyll content (CCC) retrieval and validation methods used in the present study. The prior knowledge of an adjusted average leaf angle (ALAadj) is involved in the PROSAIL-based CCC retrieval model using the random forest regression (RFR) algorithm. Full article ">Figure 2
Conceptual effect of ALAadj in (c) on bridging the difference in soil gap fraction cover between (a) a homogeneous canopy case and (b) a heterogeneous canopy case with the same leaf area index (LAI). LAI = LA/A. Full article ">Figure 3
Variable importance analysis for canopy reflectance at different wavelengths in the two cases of a homogeneous canopy and vegetation-soil mixed pixels. Full article ">Figure 4
Scatter plots of the ground-measured CCC and the CCC predicted by ground-measured canopy spectra, based on the RFR models (a–c) without and (d) with prior ALAadj knowledge related to crop types. (a) NonPrior-ALA_45 mode; (b) NonPrior-ALA_62 model; (c) NonPrior-ALA_45_62 model; (d) Prior-ALA_45_62 model. Full article ">Figure 5
Scatter plots of the ground-measured and predicted CCC based on the RFR model using continuous and discrete spectra. Above: results of NonPrior-ALA_45_62 model using (a) 50 bands, (b) 13 bands, and (c) nine bands. Below: results of Prior-ALA_45_62 model using (d) 50 bands, (e) 13 bands, and (f) nine bands. Full article ">Figure 6
Accuracy of CCC retrieval characterized by LAIs for wheat and soybean using ground-measured continuous spectra (a) without and (b) with considering prior ALAadj knowledge related to crop types. (a) NonPrior-ALA_45_62 model; (b) Prior-ALA_45_62 model. Full article ">Figure 7
Applicability of the ALAadj parameterization model for nadir observation to different view zenith angles. The test scenario is a homogeneous canopy simulated by the PROSAIL model, with variations in the view zenith angle and a fixed ALA of 45 degrees. Here, the fixed ALA is equal to the ALAadj truth value. Full article ">

21 pages, 8842 KiB

Open AccessArticle

High-Resolution Gridded Livestock Projection for Western China Based on Machine Learning

by Xianghua Li, Jinliang Hou and Chunlin Huang

Remote Sens. 2021, 13(24), 5038; https://doi.org/10.3390/rs13245038 - 11 Dec 2021

Cited by 18 | Viewed by 4428

Abstract

Accurate high-resolution gridded livestock distribution data are of great significance for the rational utilization of grassland resources, environmental impact assessment, and the sustainable development of animal husbandry. Traditional livestock distribution data are collected at the administrative unit level, which does not provide a [...] Read more.

Accurate high-resolution gridded livestock distribution data are of great significance for the rational utilization of grassland resources, environmental impact assessment, and the sustainable development of animal husbandry. Traditional livestock distribution data are collected at the administrative unit level, which does not provide a sufficiently detailed geographical description of livestock distribution. In this study, we proposed a scheme by integrating high-resolution gridded geographic data and livestock statistics through machine learning regression models to spatially disaggregate the livestock statistics data into 1 km × 1 km spatial resolution. Three machine learning models, including support vector machine (SVM), random forest (RF), and deep neural network (DNN), were constructed to represent the complex nonlinear relationship between various environmental factors (e.g., land use practice, topography, climate, and socioeconomic factors) and livestock density. By applying the proposed method, we generated a set of 1 km × 1 km spatial distribution maps of cattle and sheep for western China from 2000 to 2015 at five-year intervals. Our projected cattle and sheep distribution maps reveal the spatial heterogeneity structures and change trend of livestock distribution at the grid level from 2000 to 2015. Compared with the traditional census livestock density, the gridded livestock distribution based on DNN has the highest accuracy, with the determinant coefficient (R²) of 0.75, root mean square error (RMSE) of 9.82 heads/km² for cattle, and the R² of 0.73, RMSE of 31.38 heads/km² for sheep. The accuracy of the RF is slightly lower than the DNN but higher than the SVM. The projection accuracy of the three machine learning models is superior to those of the published Gridded Livestock of the World (GLW) datasets. Consequently, deep learning has the potential to be an effective tool for high-resolution gridded livestock projection by combining geographic and census data. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Graphical abstract

20 pages, 28343 KiB

Open AccessArticle

Estimating Vertical Distribution of Leaf Water Content within Wheat Canopies after Head Emergence

by Weiping Kong, Wenjiang Huang, Lingling Ma, Lingli Tang, Chuanrong Li, Xianfeng Zhou and Raffaele Casa

Remote Sens. 2021, 13(20), 4125; https://doi.org/10.3390/rs13204125 - 14 Oct 2021

Cited by 7 | Viewed by 2646

Abstract

Monitoring vertical profile of leaf water content (LWC) within wheat canopies after head emergence is vital significant for increasing crop yield. However, the estimation of vertical distribution of LWC from remote sensing data is still challenging due to the effects of wheat spikes [...] Read more.

Monitoring vertical profile of leaf water content (LWC) within wheat canopies after head emergence is vital significant for increasing crop yield. However, the estimation of vertical distribution of LWC from remote sensing data is still challenging due to the effects of wheat spikes and the efficacy of sensor measurement from the nadir direction. Using two-year field experiments with different growth stages after head emergence, N rates, wheat cultivars, we investigated the vertical distribution of LWC within canopies, the changes of canopy reflectance after spikes removal, the relationship between spectral indices and LWC in the upper-, middle- and bottom-layer. The interrelationship among vertical LWC were constructed, and four ratio of reflectance difference (RRD) type of indices were proposed based on the published WI and NDWSI indices to determine vertical distribution of LWC. The results indicated a bell shape distribution of LWC in wheat plants with the highest value appeared at the middle layer, and significant linear correlations between middle-LWC vs. upper-LWC and middle-LWC vs. bottom-LWC (r ? 0.92) were identified. The effects of wheat spikes on spectral reflectance mainly occurred in near infrared to shortwave infrared regions, which then decreased the accuracy of LWC estimation. Spectral indices at the middle layer outperformed the other two layers in LWC assessment and were less susceptible to wheat spikes effects, in particular, the newly proposed narrow-band WI-4 and NDWSI-4 indices exhibited great potential in tracking the changes of middle-LWC (R² = 0.82 and 0.84, respectively). By taking into account the effects of wheat spikes and the interrelationship of vertical LWC within canopies, an indirect induction strategy was developed for modeling the upper-LWC and bottom-LWC. It was found that the indirect induction models based on the WI-4 and NDWSI-4 indices were more effective than the models obtained from conventional direct estimation method, with R² of 0.78 and 0.81 for the upper-LWC estimation, and 0.75 and 0.74 for the bottom-LWC estimation, respectively. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Graphical abstract

22 pages, 6796 KiB

Open AccessArticle

Three-Dimensional Convolutional Neural Network Model for Early Detection of Pine Wilt Disease Using UAV-Based Hyperspectral Images

by Run Yu, Youqing Luo, Haonan Li, Liyuan Yang, Huaguo Huang, Linfeng Yu and Lili Ren

Remote Sens. 2021, 13(20), 4065; https://doi.org/10.3390/rs13204065 - 11 Oct 2021

Cited by 51 | Viewed by 4519

Abstract

As one of the most devastating disasters to pine forests, pine wilt disease (PWD) has caused tremendous ecological and economic losses in China. An effective way to prevent large-scale PWD outbreaks is to detect and remove the damaged pine trees at the early [...] Read more.

As one of the most devastating disasters to pine forests, pine wilt disease (PWD) has caused tremendous ecological and economic losses in China. An effective way to prevent large-scale PWD outbreaks is to detect and remove the damaged pine trees at the early stage of PWD infection. However, early infected pine trees do not show obvious changes in morphology or color in the visible wavelength range, making early detection of PWD tricky. Unmanned aerial vehicle (UAV)-based hyperspectral imagery (HI) has great potential for early detection of PWD. However, the commonly used methods, such as the two-dimensional convolutional neural network (2D-CNN), fail to simultaneously extract and fully utilize the spatial and spectral information, whereas the three-dimensional convolutional neural network (3D-CNN) is able to collect this information from raw hyperspectral data. In this paper, we applied the residual block to 3D-CNN and constructed a 3D-Res CNN model, the performance of which was then compared with that of 3D-CNN, 2D-CNN, and 2D-Res CNN in identifying PWD-infected pine trees from the hyperspectral images. The 3D-Res CNN model outperformed the other models, achieving an overall accuracy (OA) of 88.11% and an accuracy of 72.86% for detecting early infected pine trees (EIPs). Using only 20% of the training samples, the OA and EIP accuracy of 3D-Res CNN can still achieve 81.06% and 51.97%, which is superior to the state-of-the-art method in the early detection of PWD based on hyperspectral images. Collectively, 3D-Res CNN was more accurate and effective in early detection of PWD. In conclusion, 3D-Res CNN is proposed for early detection of PWD in this paper, making the prediction and control of PWD more accurate and effective. This model can also be applied to detect pine trees damaged by other diseases or insect pests in the forest. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

16 pages, 26173 KiB

Open AccessArticle

Using UAV-Based Hyperspectral Imagery to Detect Winter Wheat Fusarium Head Blight

by Huiqin Ma, Wenjiang Huang, Yingying Dong, Linyi Liu and Anting Guo

Remote Sens. 2021, 13(15), 3024; https://doi.org/10.3390/rs13153024 - 1 Aug 2021

Cited by 37 | Viewed by 4786

Abstract

Fusarium head blight (FHB) is a major winter wheat disease in China. The accurate and timely detection of wheat FHB is vital to scientific field management. By combining three types of spectral features, namely, spectral bands (SBs), vegetation indices (VIs), and wavelet features [...] Read more.

Fusarium head blight (FHB) is a major winter wheat disease in China. The accurate and timely detection of wheat FHB is vital to scientific field management. By combining three types of spectral features, namely, spectral bands (SBs), vegetation indices (VIs), and wavelet features (WFs), in this study, we explore the potential of using hyperspectral imagery obtained from an unmanned aerial vehicle (UAV), to detect wheat FHB. First, during the wheat filling period, two UAV-based hyperspectral images were acquired. SBs, VIs, and WFs that were sensitive to wheat FHB were extracted and optimized from the two images. Subsequently, a field-scale wheat FHB detection model was formulated, based on the optimal spectral feature combination of SBs, VIs, and WFs (SBs + VIs + WFs), using a support vector machine. Two commonly used data normalization algorithms were utilized before the construction of the model. The single WFs, and the spectral feature combination of optimal SBs and VIs (SBs + VIs), were respectively used to formulate models for comparison and testing. The results showed that the detection model based on the normalized SBs + VIs + WFs, using min–max normalization algorithm, achieved the highest R² of 0.88 and the lowest RMSE of 2.68% among the three models. Our results suggest that UAV-based hyperspectral imaging technology is promising for the field-scale detection of wheat FHB. Combining traditional SBs and VIs with WFs can improve the detection accuracy of wheat FHB effectively. Full article

(This article belongs to the Special Issue The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing)

► Show Figures

Figure 1

Figure 1
Experimental area location and field survey sampling positions. Full article ">Figure 2
Methodological framework of wheat FHB detection model. Full article ">Figure 3
Disease-sensitive wavelet regions extracted from UAV hyperspectral imagery. Full article ">Figure 4
R2 among different features of each spectral feature type: (a) SBs, (b) VIs, and (c) WFs. Full article ">Figure 4 Cont.
R2 among different features of each spectral feature type: (a) SBs, (b) VIs, and (c) WFs. Full article ">Figure 5
Scatter plots of measured DER versus estimated DER of SVM detection models based on (a–c) initial spectral feature combinations of SBs, VIs, and WFs; (d–f) spectral feature combinations of SBs, VIs, and WFs normalized using MMN algorithm; and (g–i) spectral feature combinations of SBs, VIs, and WFs normalized using ZSN algorithm. Full article ">Figure 6
Damage maps of wheat FHB on (a) 3 May and (b) 8 May 2019, obtained using the normalized (SBs + VIs + WFs)-based SVM detection model. Full article ">Figure 6 Cont.
Damage maps of wheat FHB on (a) 3 May and (b) 8 May 2019, obtained using the normalized (SBs + VIs + WFs)-based SVM detection model. Full article ">

Journal Menu

Journal Browser

The Emerging Trends and Applications of Big Data and Machine Learning/Artificial Intelligence (AI) in Remote Sensing

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Related Special Issue

Published Papers (13 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI