Estimation of Building Density with the Integrated Use of GF-1 PMS and Radarsat-2 Data
"> Figure 1
<p>(<b>left</b>) The geographical location; and (<b>right</b>) the GF-1 PMS images of the studied regions (areas in the red square are used for training and validation).</p> "> Figure 2
<p>Flowchart of building density modeling and mapping.</p> "> Figure 3
<p>Principle of pixel aggregate method: (<b>a</b>) input image; (<b>b</b>) output image; and (<b>c</b>) overlap image.</p> "> Figure 4
<p>Samples with similar density and different distribution pattern: (<b>a</b>) building density: 42.43%, intensity: −15.81; (<b>b</b>) building density: 42.22%, intensity: −13.33; (<b>c</b>) building density: 44.38%, intensity: −15.33; (<b>d</b>) building density: 42.02%, intensity: −9.86; (<b>e</b>) building density: 43.28%, intensity: −16.91; and (<b>f</b>) building density: 44.71%, intensity: −16.24.</p> "> Figure 5
<p>Triangular prism method: (<b>a</b>) Top view of the corner pixels and the center point used in improved triangular prism method (an example with step size = 4); and (<b>b</b>) 3D view of improved triangular prism method.</p> "> Figure 6
<p>Illustration of the reiteration procedure to calculate top surface area. (<b>a</b>) <math display="inline"> <semantics> <mrow> <mi>s</mi> <mi>t</mi> <mi>e</mi> <mi>p</mi> <mi>s</mi> <mi>i</mi> <mi>z</mi> <mi>e</mi> <mo>=</mo> <mn>2</mn> <mo>,</mo> <mtext> </mtext> <mi>t</mi> <mi>o</mi> <mi>p</mi> <mtext> </mtext> <mi>s</mi> <mi>u</mi> <mi>r</mi> <mi>f</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mtext> </mtext> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>a</mi> <mo>=</mo> <mstyle displaystyle="true"> <mo>∑</mo> <mrow> <mrow> <mo>(</mo> <mrow> <mn>64</mn> <mo>×</mo> <mi>s</mi> <mi>q</mi> <mi>u</mi> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>s</mi> </mrow> <mo>)</mo> </mrow> </mrow> </mstyle> </mrow> </semantics> </math>; (<b>b</b>) <math display="inline"> <semantics> <mrow> <mi>s</mi> <mi>t</mi> <mi>e</mi> <mi>p</mi> <mi>s</mi> <mi>i</mi> <mi>z</mi> <mi>e</mi> <mo>=</mo> <mn>4</mn> <mo>,</mo> <mtext> </mtext> <mi>t</mi> <mi>o</mi> <mi>p</mi> <mtext> </mtext> <mi>s</mi> <mi>u</mi> <mi>r</mi> <mi>f</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mtext> </mtext> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>a</mi> <mo>=</mo> <mstyle displaystyle="true"> <mo>∑</mo> <mrow> <mrow> <mo>(</mo> <mrow> <mn>16</mn> <mo>×</mo> <mi>s</mi> <mi>q</mi> <mi>u</mi> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>s</mi> </mrow> <mo>)</mo> </mrow> </mrow> </mstyle> </mrow> </semantics> </math>; (<b>c</b>) <math display="inline"> <semantics> <mrow> <mi>s</mi> <mi>t</mi> <mi>e</mi> <mi>p</mi> <mi>s</mi> <mi>i</mi> <mi>z</mi> <mi>e</mi> <mo>=</mo> <mn>8</mn> <mo>,</mo> <mtext> </mtext> <mi>t</mi> <mi>o</mi> <mi>p</mi> <mtext> </mtext> <mi>s</mi> <mi>u</mi> <mi>r</mi> <mi>f</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mtext> </mtext> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>a</mi> <mo>=</mo> <mstyle displaystyle="true"> <mo>∑</mo> <mrow> <mrow> <mo>(</mo> <mrow> <mn>4</mn> <mo>×</mo> <mi>s</mi> <mi>q</mi> <mi>u</mi> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>s</mi> </mrow> <mo>)</mo> </mrow> </mrow> </mstyle> </mrow> </semantics> </math>; and (<b>d</b>) <math display="inline"> <semantics> <mrow> <mi>s</mi> <mi>t</mi> <mi>e</mi> <mi>p</mi> <mi>s</mi> <mi>i</mi> <mi>z</mi> <mi>e</mi> <mo>=</mo> <mn>16</mn> <mo>,</mo> <mtext> </mtext> <mi>t</mi> <mi>o</mi> <mi>p</mi> <mtext> </mtext> <mi>s</mi> <mi>u</mi> <mi>r</mi> <mi>f</mi> <mi>a</mi> <mi>c</mi> <mi>e</mi> <mtext> </mtext> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>a</mi> <mo>=</mo> <mstyle displaystyle="true"> <mo>∑</mo> <mrow> <mrow> <mo>(</mo> <mrow> <mn>1</mn> <mo>×</mo> <mi>s</mi> <mi>q</mi> <mi>u</mi> <mi>a</mi> <mi>r</mi> <mi>e</mi> <mi>s</mi> </mrow> <mo>)</mo> </mrow> </mrow> </mstyle> </mrow> </semantics> </math>.</p> "> Figure 7
<p>Relative differential box counting method (in the figure, the moving window size is 9 × 9 and the gliding box size is 3).</p> "> Figure 8
<p>Distribution pattern of different building areas from Google Map images ((<b>a</b>–<b>l</b>) represent different regions in the study area).</p> "> Figure 9
<p>Distribution of Deviation Degree (DD): (<b>a</b>) Jizhou; and (<b>b</b>) Beijing.</p> "> Figure 10
<p>Fractal dimensions and lacunarities of samples: (<b>a</b>) fractal dimension; and (<b>b</b>) lacunarity.</p> "> Figure 11
<p>(<b>a</b>) Backscatter intensity (<span class="html-italic">BI</span>) distribution of Jizhou; (<b>b</b>) amended backscatter intensity (<span class="html-italic">ABI</span>) distribution of Jizhou; (<b>c</b>) backscatter intensity (<span class="html-italic">BI</span>) distribution of Beijing; and (<b>d</b>) amended backscatter intensity (<span class="html-italic">ABI</span>) distribution of Beijing.</p> "> Figure 12
<p>Overall building density distribution of Jizhou: (<b>a</b>) model <span class="html-italic">a</span> using four GF-1 PMS spectral bands (<span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>); (<b>b</b>) model <span class="html-italic">b</span> using Normalized Difference Vegetation Index (<span class="html-italic">NDVI</span>), Normalized Difference Water Index (<span class="html-italic">NDWI</span>), and Ratio Built-up Index (<span class="html-italic">RBI</span>); (<b>c</b>) model <span class="html-italic">c</span> using <span class="html-italic">BI</span>; (<b>d</b>) model <span class="html-italic">d</span> using <span class="html-italic">ABI</span>; (<b>e</b>) model <span class="html-italic">e</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDBI</span>, and <span class="html-italic">RBI</span>; (<b>f</b>) model <span class="html-italic">f</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub> and <span class="html-italic">BI</span>; (<b>g</b>) model <span class="html-italic">g</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub> and <span class="html-italic">ABI</span>; (<b>h</b>) model <span class="html-italic">h</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDWI</span>, <span class="html-italic">RBI</span>, and <span class="html-italic">BI</span>; and (<b>i</b>) model <span class="html-italic">i</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDWI</span>, <span class="html-italic">RBI</span>, and <span class="html-italic">ABI</span>.</p> "> Figure 13
<p>Overall building density distribution of Beijing: (<b>a</b>) model <span class="html-italic">a</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>; (<b>b</b>) model <span class="html-italic">b</span> using <span class="html-italic">NDVI</span>, <span class="html-italic">NDBI</span>, and <span class="html-italic">RBI</span>; (<b>c</b>) model <span class="html-italic">c</span> using <span class="html-italic">BI</span>; (<b>d</b>) model <span class="html-italic">d</span> using <span class="html-italic">ABI</span>; (<b>e</b>) model <span class="html-italic">e</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDBI</span>, and <span class="html-italic">RBI</span>; (<b>f</b>) model <span class="html-italic">f</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub> and <span class="html-italic">BI</span>; (<b>g</b>) model <span class="html-italic">g</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub> and <span class="html-italic">ABI</span>; (<b>h</b>) model <span class="html-italic">h</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDWI</span>, <span class="html-italic">RBI</span>, and <span class="html-italic">BI</span>; and (<b>i</b>) model <span class="html-italic">i</span> using <span class="html-italic">B</span><sub>1<span class="html-italic">–</span>4</sub>, <span class="html-italic">NDVI</span>, <span class="html-italic">NDWI</span>, <span class="html-italic">RBI</span>, and <span class="html-italic">ABI</span>.</p> ">
Abstract
:1. Introduction
2. Study Area and Data
2.1. Study Area
2.2. Data
3. Method
3.1. Sample and Feature Selection
3.2. Amended Backscatter Intensity
3.2.1. Fractal Dimension and Lacunarity
3.2.2. Deviation Degree
3.2.3. Calculation of Amended Backscatter Intensity
3.3. Accuracy Assessment
4. Result
4.1. Distribution of Deviation Degree
4.2. Establishment of Decision Rules
4.3. Distribution of Amended Backscatter Intensity
4.4. Comparison between Different Models
5. Discussion
- When generating ABI, we only cared about the influence caused by spatial pattern, but ignored height, building orientation, shadow and layover. To sum up, it is a 2D-based method without consideration of 3D information.
- The 8 m resolution GF-1 image is relatively coarse and may lose some detailed information when calculating fractal dimension and lacunarity.
- In determining the coefficient of DD, we have developed a series of decision rules to classify the image into some types. However, the selection of class varies from city to city and the coefficient determination need extensive analysis. Therefore, a unified classification scheme and an automatic method for determining coefficient are urgently needed.
- Testing samples were manually selected, and although randomness was emphasized during the selection, a small number of samples is insufficient for representing the entire region.
- The study area is a plain where the landscapes are relatively simple (e.g., the regularly-shaped, and equally-scaled farmland). However, when the method is extended to a hilly area, it may not be that precise since the natural landscapes will be more complex and their landscape metrics may be difficult to distinguish with building areas.
- The problem for seasonality persists. Although the result showed that the calculations of fractal dimension, lacunarity, and deviation degree are not affected, we cannot be certain whether the overall higher backscatter intensities are caused by seasonality or incidence angle or building types. Therefore how different seasons will affect the SAR backscattering pattern needs to be addressed in the future.
6. Conclusions
Acknowledgments
Author Contributions
Conflicts of Interest
References
- Wu, C.; Murray, A.T. Population estimation using Landsat enhanced thematic mapper imagery. Geogr. Anal. 2007, 39, 26–43. [Google Scholar] [CrossRef]
- Zeng, C. Spatial and Temporal Analysis of Population Distribution in China Based on Remote Sensing Data. Master’s Thesis, Institute of Remote Sensing Applications, Chinese Acamedy of Sciences, Beijing, China, 2010. [Google Scholar]
- Azar, D.; Graesser, J.; Engstrom, R.; Comenetz, J.; Leddy, R.M.; Schechtman, N.G.; Andrews, T. Spatial refinement of census population distribution using remotely sensed estimates of impervious surfaces in haiti. Int. J. Remote Sens. 2010, 31, 5635–5655. [Google Scholar] [CrossRef]
- Sawaya, K.E.; Olmanson, L.G.; Heinert, N.J.; Brezonik, P.L.; Bauer, M.E. Extending satellite remote sensing to local scales: Land and water resource monitoring using high-resolution imagery. Remote Sens. Environ. 2003, 1, 144–156. [Google Scholar] [CrossRef]
- Bauer, M.E.; Loffelholz, B.C.; Wilson, B.; Loeffelholz, B.C. Estimating and mapping impervious surface area by regression analysis of landsat imagery. In Remote Sensing of Impervious Surfaces; Weng, Q., Ed.; CRC Press: Boca Raton, FL, USA, 2008; pp. 3–19. [Google Scholar]
- Yang, L.; Huang, C.; Homer, C.G.; Wylie, B.K.; Coan, M.J. An approach for mapping large-area impervious surfaces: Synergistic use of Landsat-7 ETM+ and high spatial resolution imagery. Can. J. Remote Sens. 2003, 29, 230–240. [Google Scholar] [CrossRef]
- Azar, D.; Engstrom, R.; Graesser, J.; Comenetz, J. Generation of fine-scale population layers using multi-resolution satellite imagery and geospatial data. Remote Sens. Environ. 2013, 130, 219–232. [Google Scholar] [CrossRef]
- Yue, W.; Wu, C. Impervious surface distribution estimation by spectral mixture analysis. J. Remote Sens. 2007, 11, 914–922. [Google Scholar]
- Fan, F.; Fan, W.; Weng, Q. Improving urban impervious surface mapping by linear spectral mixture analysis and using spectral indices. Can. J. Remote Sens. 2015, 41. [Google Scholar] [CrossRef]
- Zhang, Y.; Zhang, H.; Lin, H. Improving the impervious surface estimation with combined use of optical and SAR remote sensing images. Remote Sens. Environ. 2014, 141, 155–167. [Google Scholar] [CrossRef]
- Yang, L.; Jiang, L.; Lin, H.; Liao, M. Quantifying sub-pixel urban impervious surface through fusion of optical and insar imagery. GISci. Remote Sens. 2009, 46, 161–171. [Google Scholar] [CrossRef]
- Lu, D.; Batistella, M.; Moran, E. Land cover classification in the brazilian amazon with the integration of Landsat ETM+ and radarsat data. Int. J. Remote Sens. 2007, 28, 5447–5459. [Google Scholar] [CrossRef]
- Leinenkugel, P.; Esch, T.; Kuenzer, C. Settlement detection and impervious surface estimation in the mekong delta using optical and SAR remote sensing data. Remote Sens. Environ. 2011, 115, 3007–3019. [Google Scholar] [CrossRef]
- Leinenkugel, P. The Combined Use of Optical and SAR Data for Large Area Impervious Surface Mapping. Master’s Thesis, University Salzburg, Salzburg, Austria, 2010. [Google Scholar]
- Zhang, J.; Yang, J.; Zhao, Z.; Li, H.; Zhang, Y. Block-regression based fusion of optical and SAR imagery for feature enhancement. Int. J. Remote Sens. 2010, 31, 2325–2345. [Google Scholar] [CrossRef]
- Zhang, H.; Zhang, Y.; Lin, H. A comparison study of impervious surfaces estimation using optical and SAR remote sensing images. Int. J. Appl. Earth Obs. Geoinform. 2012, 18, 148–156. [Google Scholar] [CrossRef]
- Slonecker, E.T. Remote sensing of impervious surfaces: A review. Remote Sens. Rev. 2001, 20, 227–255. [Google Scholar] [CrossRef]
- Schueler, T. The importance of imperviousness. Watershed Prot. Tech. 1994, 1, 100–111. [Google Scholar]
- Li, J.; Zhang, L.; Wu, B.; Ma, X. Study on extracting building density and floor area ratio based on high resolution image. Remote Sens. Technol. Appl. 2007, 3, 309–313. [Google Scholar]
- Yu, B.; Liu, H.; Wu, J.; Hu, Y.; Zhang, L. Automated derivation of urban building density information using airborne LiDAR data and object-based method. Landsc. Urban Plan. 2010, 98, 210–219. [Google Scholar] [CrossRef]
- Wu, Q.; Chen, R.; Sun, H.; Cao, Y. Urban building density detection using high resolution SAR imagery. In Proceedings of the 2011 Urban Remote Sensing Event (JURSE), Munich, Germany, 11–13 April 2011.
- Cao, Y.; Su, C.; Liang, J. Building unit density detection from high resolution TerraSAR-X image based on mathematical morphological operators. In Proceedings of the 2012 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Munich, Germany, 22–27 July 2012.
- Kajimoto, M.; Susaki, J. Urban-area extraction from polarimetric sar images using polarization orientation angle. In Proceedings of the 2013 Geoscience and Remote Sensing Symposium, Melbourne, Australia, 21–26 July 2013.
- Susaki, J.; Kajimoto, M.; Kishimoto, M.; Susaki, J.; Kajimoto, M.; Kishimoto, M. Urban density mapping of global megacities from polarimetric SAR images. Remote Sens. Environ. 2014, 155, 334–348. [Google Scholar] [CrossRef] [Green Version]
- Stasolla, M.; Gamba, P. Spatial indexes for the extraction of formal and informal human settlements from high-resolution sar images. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2008, 1, 98–106. [Google Scholar] [CrossRef]
- Zheng, H.; Zeng, Y.; Wang, Y.; Sheng, S.; Qiu, X. Urban building density estimation based on VIIRS night-time satellite data-a case of Nanjing. Sci. Technol. Eng. 2014, 18, 68–75. [Google Scholar]
- Wu, W.; Zhao, S.; Zhu, C.; Jiang, J. A comparative study of urban expansion in Beijing, Tianjin and Shijiazhuang over the past three decades. Landsc. Urban Plan. 2015, 134, 93–106. [Google Scholar] [CrossRef]
- Franceschetti, G.; Iodice, A.; Riccio, D. A canonical problem in electromagnetic backscattering from buildings. IEEE Trans. Geosci. Remote 2002, 40, 1787–1801. [Google Scholar] [CrossRef]
- Breiman, L.; Friedman, J.H.; Olshen, R.A.; Stone, C.J. Classification and Regression Trees; Wadsworth International Group: Belmont, CA, USA, 1984. [Google Scholar]
- Weng, Q. Remote sensing of impervious surfaces in the urban areas: Requirements, methods, and trends. Remote Sens. Environ. 2012, 117, 34–49. [Google Scholar] [CrossRef]
- Huang, C.; Townshend, J.R.G. A stepwise regression tree for nonlinear approximation: Applications to estimating subpixel land cover. Int. J. Remote Sens. 2003, 24, 75–90. [Google Scholar] [CrossRef]
- Yang, W.; Zhang, Y.; Yin, X.; Wang, J. Construction of ratio built-up index for GF-1 image. Remote Sens. Land. Resour. 2016, 28, 35–42. [Google Scholar]
- Horne, J.H.; Horne, J.H. A tasseled cap transformation for IKONOS images. In Proceedings of the ASPRS 2003 Annual Conference, Anchorage, AK, USA, 4–5 May 2003.
- Zhao, L. Building Extraction from High-Resolution SAR Imagery. Ph.D. Thesis, National University of Defense Technology, Changsha, China, 2009. [Google Scholar]
- Schmidt, M.; Esch, T.; Klein, D.; Thiel, M.; Dech, S. Estimation of building density using TerraSAR-X-data. In Proceedings of the 2010 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Honolulu, HI, USA, 25–30 July 2010; pp. 1936–1939.
- Sun, W.; Xu, G.; Gong, P.; Liang, S. Fractal analysis of remotely sensed images: A review of methods and applications. Int. J. Remote Sens. 2006, 27, 4963–4990. [Google Scholar] [CrossRef]
- Myint, S.W.; Lam, N. Examining lacunarity approaches in comparison with fractal and spatial autocorrelation techniques for urban mapping. Photogramm. Eng. Remote Sens. 2005, 71, 927–937. [Google Scholar] [CrossRef]
- Smde, J.; Burroogh, P.A. A fractal approach to the classification of Mediterranean vegetation types in remotely sensed data. Photogramm. Eng. Remote Sens. 1995, 61, 1041–1053. [Google Scholar]
- Qiu, H.I.; Lam, S.N.; Quattrochi, D.A.; Gamon, J.A.; Qiu, H.I.; Gamon, J.A. Fractal characterization of hyperspectral imagery. Photogramm. Eng. Remote Sens. 1999, 65, 63–71. [Google Scholar]
- Lam, N.S.-N.; Qiu, H.-L.; Quattrochi, D.A.; Emerson, C.W. An evaluation of fractal methods for characterizing image complexity. Cartogr. Geogr. Inf. Sci. 2002, 29, 25–35. [Google Scholar] [CrossRef]
- Mandelbrot, B.B. Fractals: Form, chance and dimension. Phys. Today 1979, 1, 725–729. [Google Scholar] [CrossRef]
- Sun, W. Three new implementations of the triangular prism method for computing the fractal dimension of remote sensing images. Photogramm. Eng. Remote Sens. 2006, 72, 373–382. [Google Scholar] [CrossRef]
- Myint, S.W.; Lam, N. A study of lacunarity-based texture analysis approaches to improve urban image classification. Comput. Environ. Urban Syst. 2005, 29, 501–523. [Google Scholar] [CrossRef]
- Dong, P. Test of a new lacunarity estimation method for image texture analysis. Int. J. Remote Sens. 2000, 21, 3369–3373. [Google Scholar] [CrossRef]
- Voss, R.; Peitgen, H.O.; Saupe, D. The Science of Fractal Images; Springer: New York, NY, USA, 1988. [Google Scholar]
- Dong, P. Lacunarity for spatial heterogeneity measurement in GIS. Geogr. Inf. Sci. 2000, 6, 20–26. [Google Scholar] [CrossRef]
- Gan, D.; Tat Soon, Y. A novel lacunarity estimation method applied to SAR image segmentation. IEEE Trans. Geosci. Remote 2002, 40, 2687–2691. [Google Scholar] [CrossRef]
- Backes, A.R. A new approach to estimate lacunarity of texture images. Pattern Recogn. Lett. 2013, 34, 1455–1461. [Google Scholar] [CrossRef]
- Michie, D.; Spiegelhalter, D.J.; Taylor, C.C.; Campbell, J. Machine Learning, Neural and Statistical Classification; Ellis Horwood Ltd.: New York, NY, USA, 1994; p. 289. [Google Scholar]
Satellite | Path | Row | Obtained Date |
---|---|---|---|
Radarsat-2 | \ | \ | 25 April 2015 |
Radarsat-2 | \ | \ | 19 February 2015 |
GF-1 | 2 | 92 | 16 August 2015 |
GF-1 | 2 | 93 | 16 August 2015 |
GF-1 | 2 | 92 | 28 August 2015 |
GF-1 | 2 | 92 | 12 October 2015 |
GF-1 | 2 | 93 | 12 October 2015 |
GF-1 | 1 | 89 | 22 January 2015 |
GF-1 | 1 | 89 | 03 January 2015 |
GF-1 | 1 | 89 | 15 February 2015 |
GF-1 | 2 | 89 | 22 January 2015 |
GF-1 | 2 | 89 | 03 January 2015 |
Google Map | \ | \ | 11 March 2015 |
Google Map | \ | \ | 07 August 2015 |
Samples | Building Density (%) | Intensity(dB) | Fractal Dimension | Lacunarity |
---|---|---|---|---|
a | 42.43 | −15.81 | 2.4638 | 1.1056 |
b | 42.22 | −13.33 | 2.5566 | 1.1209 |
c | 44.38 | −15.33 | 2.6199 | 1.1536 |
d | 42.02 | −9.86 | 2.4378 | 1.1441 |
e | 43.28 | −16.91 | 2.1137 | 1.4574 |
f | 44.71 | −16.24 | 2.4821 | 1.1578 |
Samples | Fractal Dimension | Lacunarity | Deviation Degree |
---|---|---|---|
a | 2.4541 | 1.1461 | 0.3460 |
b | 2.5445 | 1.0853 | 0.2704 |
c | 2.2360 | 1.3400 | 0.5520 |
d | 2.4881 | 1.3036 | 0.4078 |
e | 2.1683 | 1.3282 | 0.5799 |
f | 2.2017 | 1.2906 | 0.5445 |
g | 2.4811 | 1.1354 | 0.3272 |
h | 2.1176 | 2.0000 | 0.9412 |
i | 2.0015 | 1.5323 | 0.7654 |
j | 2.4283 | 1.1529 | 0.3623 |
k | 2.5583 | 1.1158 | 0.2788 |
l | 2.5565 | 1.1898 | 0.3166 |
Jizhou | Beijing | ||||
---|---|---|---|---|---|
Class | Characteristics | Coefficient | Class | Characteristic | Coefficient |
Non-built-up Area (0%) | Backscatter intensity ; | 1 | Non-built-up Area (0%) | Backscatter intensity ; | 1 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
Water (0%) | Backscatter intensity ; | 0 | Water (0%) | Backscatter intensity ; | 0 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
Low density area (14.29%) | Backscatter intensity ; | 1 | Workshop-like function area (60.40%) | Backscatter intensity ; | −1 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
Middle density area (29.88%) | Backscatter intensity ; | 1 | High density level community (27.53%) | Backscatter intensity ; | −1 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
High density area (36.26%) | Backscatter intensity ; | −1 | Low density level community (10.12%) | Backscatter intensity ; | 0 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
Very high density area (50.34%) | Backscatter intensity ; | −1 | Industrial area (64.05%) | Backscatter intensity ; | −1 |
Fractal dimension ; | Fractal dimension ; | ||||
Lacunarity ; | Lacunarity ; | ||||
Deviation degree | Deviation degree | ||||
\ | \ | \ | Non-residential built-up area (17.96%) | Backscatter intensity ; | −1 |
Fractal dimension ; | |||||
Lacunarity ; | |||||
Deviation degree |
Jizhou | Beijing | ||||
---|---|---|---|---|---|
Class | Produce Accuracy (%) | User Accuracy (%) | Class | Produce Accuracy (%) | User Accuracy (%) |
Non-built-up area | 81.25 | 81.25 | Non-built-up area | 78.57 | 86.42 |
Water | 94.12 | 88.89 | Water | 85.71 | 92.31 |
Low density area | 70.59 | 66.67 | Workshop-like function area | 60.00 | 56.25 |
Middle density area | 50.00 | 47.06 | High density level community | 64.28 | 90.00 |
High density area | 58.82 | 71.43 | Low density level community | 53.33 | 44.44 |
Very high density area | 70.59 | 70.59 | Industrial area | 78.57 | 64.71 |
\ | \ | Non-residential built-up Area | 50.00 | 53.85 | |
Overall accuracy | 71.00 | Overall accuracy | 67.00 |
Models | Input Features |
---|---|
a | B1–4 |
b | NDVI, NDWI, RBI |
c | BI |
d | ABI |
e | B1–4, NDVI, NDWI, RBI |
f | B1–4, BI |
g | B1–4, ABI |
h | B1–4, NDVI, NDWI, RBI, BI |
i | B1–4, NDVI, NDWI, RBI, ABI |
Model | Jizhou | Beijing | ||||
---|---|---|---|---|---|---|
RMSE | R2 | p | RMSE | R2 | p | |
a | 15.45 | 0.35 | 0.01/0.005/0.025 | 21.78 | 0.26 | 0.025/0.01 |
b | 16.57 | 0.43 | 0.005 | 24.23 | 0.19 | 0.025 |
c | 12.71 | 0.48 | 0.025 | 21.56 | 0.24 | 0.005 |
d | 11.53 | 0.58 | 0.005 | 20.34 | 0.30 | 0.01/0.005 |
e | 10.28 | 0.63 | 0.005 | 20.34 | 0.36 | 0.05/0.025 |
f | 10.33 | 0.60 | 0.005/0.025 | 18.29 | 0.49 | 0.005/0.01/0.005 |
g | 9.54 | 0.77 | 0.005 | 16.91 | 0.54 | 0.005/0.005 |
h | 10.73 | 0.60 | 0.005/0.01/0.005 | 17.58 | 0.52 | 0.025/0.005/0.025 |
i | 8.93 | 0.80 | 0.005 | 16.21 | 0.64 | 0.05/0.025 |
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Zhou, Y.; Lin, C.; Wang, S.; Liu, W.; Tian, Y. Estimation of Building Density with the Integrated Use of GF-1 PMS and Radarsat-2 Data. Remote Sens. 2016, 8, 969. https://doi.org/10.3390/rs8110969
Zhou Y, Lin C, Wang S, Liu W, Tian Y. Estimation of Building Density with the Integrated Use of GF-1 PMS and Radarsat-2 Data. Remote Sensing. 2016; 8(11):969. https://doi.org/10.3390/rs8110969
Chicago/Turabian StyleZhou, Yi, Chenxi Lin, Shixin Wang, Wenliang Liu, and Ye Tian. 2016. "Estimation of Building Density with the Integrated Use of GF-1 PMS and Radarsat-2 Data" Remote Sensing 8, no. 11: 969. https://doi.org/10.3390/rs8110969