(PDF) A Method for Built-Up Area Extraction Using Dual Polarimetric Alos Palsar

ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume IV-5, 2018 ISPRS TC V Mid-term Symposium “Geospatial Technology – Pixel to People”, 20–23 November 2018, Dehradun, India A METHOD FOR BUILT-UP AREA EXTRACTION USING DUAL POLARIMETRIC ALOS PALSAR S. Sinha 1, *, A. Santra 1, S. S. Mitra 1 1 Department of Civil Engineering, Haldia Institute of Technology, Haldia , India -

Commission V, SS: Infrastructure and Development Planning KEY WORDS: Built-up Area, Synthetic Aperture Radar, ALOS PALSAR, Backscatter, Classification ABSTRACT: Mapping of the built-up area is a task of exigency as the area supports varieties of anthropogenic activities and is important for several ecosystem services. The task becomes more complicated owing to similarities in spectral characteristics with the bare soil. A new radar-based approach is proposed using Synthetic Aperture Radar (SAR) HH/HV dual polarized L-band ALOS PALSAR data. Sigma nought (σ°) values are extracted from the HH and HV polarized data. HH and HV sigma nought images are ratioed and normalized and then equated together in a unique combination to generate an index that is developed, cross-checked and validated over multiple regions, with simultaneous inputs from Ground truth (GT) data. Maps developed using the index is classified into built-up and non built-up areas, where the results show that the proposed method is very effective for built-up area detection. The approach adopted in this study is acceptable due to its high accuracy, simplicity and reliability; and hence easy to replicate. 1. INTRODUCTION optical data. SAR data from Radarsat has been used for urban land-cover mapping with high classification accuracy (Hu and Rapid globalization leads to urbanization. Several problems Ban, 2008). Contextual information from SAR data like originated due to rapid urbanization, most of which are Radarsat-2 and TerraSAR-X were used for urban land use irreversible. The increasing rate of urbanization and in-turn the mapping with moderate classification accuracy (Chen et al., built-up areas results in an interconnected consequence of 2013; Lv et al., 2015). Texture measures have been found to be environmental issues, with decrease of vegetated areas and the beneficial for extracting information related to built-up features expansion of urban heat island effects, affecting the overall (Corbane et al., 2008; Aghababaee et al., 2013; Shao et al., ecosystem and biodiversity. This has resulted in investigation 2016). Polarimetric capabilities of SAR data from Sentinel-1 concerning identification and extraction of built-up areas for were investigated for land cover mapping in urban areas, where urban planning, climate studies and resource management using the use of dual polarization proved superior to single satellite data (Bramhe et al., 2018). polarization (Abdikan et al., 2016). Classification algorithms like Maximum likelihood and K-means are effective for Remote sensing is used to establish considerable assistance in classifying segmented images generated from SAR data (Wang the analysis of urban ecosystems through objective and et al., 2016). Data fusion technique is also explored for verifiable characterization of urban composition. Mapping of delineation of urban impervious surface from the fused data land features began with the conventional multi-spectral product (Shao et al., 2016). Integrated use of optical and SAR analysis; however, with satisfactory level of classification has also been explored for extraction of urban and built-up accuracy owing to spectral mixture of heterogeneous land areas (Corbane et al., 2008; Qin et al., 2017). Synergic use of features Several studies have used optical spectral both optical and SAR surmounts the mutual constraints and characteristics in remote sensing to detect built-up areas (Sinha results in improved analysis in most cases (Sinha et al., 2016). et al., 2015a, 2018); but the heterogeneity in the satellite images results in complex patterns in the images that are complex to In this paper, we introduce a SAR backscatter based index that understand, which make the retrieval of accurate information can automatically extract built-up areas from SAR imagery. The from images even more intricate and challenging. Built-up co-polarization backscatter show a unique pattern over built-up features share similar spectral characteristics with the bare soil, areas than the cross-polarizations (Liu, 2016), hence, ratio river sands and fallow lands. Hence, spectral properties are not and/or normalization between the co- and cross-polarizations adequate to discriminate built-up areas. In order to overcome has the capability to distinguish built-up areas from other land such intricacy, radar or SAR (Synthetic Aperture Radar) remote use land cover (LULC) features. Backscatter or sigma nought sensing are considered as the best alternative, as the target- values from HH and HV polarizations are combined in a unique signal interaction rely on the scattering properties of the ground fashion to generate the index and is applied over multiple site to features, depending on which backscatter values are generated establish its potentiality and validity. The approach can easily (Sinha et al., 2015b). be implemented over various regions as it can work independently without using training samples or self-defined SAR has been applied to extract built-up information over thresholds. various scales in somewhat limited studies, either singly or with * Corresponding author This contribution has been peer-reviewed. The double-blind peer-review was conducted on the basis of the full paper. https://doi.org/10.5194/isprs-annals-IV-5-455-2018 | © Authors 2018. CC BY 4.0 License. 455 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume IV-5, 2018 ISPRS TC V Mid-term Symposium “Geospatial Technology – Pixel to People”, 20–23 November 2018, Dehradun, India 2. STUDY AREA AND DATA USED 3. METHODOLOGY The study had been carried out over multiple sites in the Indian 3.1 SAR data processing subcontinent as depicted in Figure 1. Site 1: South Kolkata, West Bengal. Megacity, high population Raw SAR data procured as single-look complex (SLC) data are density. Coordinates: 22.57oN and 88.36oE. multi-looked to convert the complex data to real numbered Site 2: Haldia, West Bengal. Industrial, moderately low intensity data. Azimuth to range ratio of 3:1 is applied to population density. Coordinates: 22.03oN and 88.06oE. generate images with squared pixels. The intensity images are Site 3: Munger, Bihar. Moderately low population density. converted to power or decibel images by radiometric calibration Coordinates: 25.37oN and 86.47oE. using Equation 1 (Sinha et al., 2016). Speckle reduction is Site 4: Jamalpur, Bihar. Very low population density. performed using Gamma map filters. Multi-looking also help in Coordinates: 25.31oN and 86.49oE. speckle reduction. Geocoding is done using orbital parameters Munger-Jamalpur is known as the twin city. with SRTM DEM resampled to 25m pixel size by nearest neighbourhood algorithm and re-projected to UTM-WGS84 coordinate system using the Range-Doppler Approach for terrain correction.    10  a log 10( DN )  Ao (1) where, σ0 is the backscatter coefficient or sigma nought values in decibels (dB), DN is the power (or intensity) image, A0 = - 115 dB is the calibration factor for ALOS PALSAR. 3.2 SAR-based index HH backscatter (HHσ0) and HV backscatter (HVσ0) values are equated to obtain ratioed (a) and normalized (b) images. Variability is observed between ea and eb for built-up features and so, the exponential products are further normalized. It resulted in a unique pattern that became even more prominent when combined together in an exponential equation, expressed as Equation 2 that helped distinguish the built-up feature class entirely from the other remaining feature classes, like water, vegetation and bare soil, specifically after when the exponential normalized product ‘(ea-eb)/(ea+eb)’ was subtracted from the normalized product ‘a’ to reduce the urban class overlapping with the water class. Built-up Index = a-c (2) where, a = [(HHσ0- HVσ0)/(HHσ0+ HVσ0)], b = (HHσ0/HVσ0), c = (ea-eb)/(ea+eb), e = exponential function. Figure 1. Location of the study sites (1. South Kolkata, 2. 4. RESULTS AND ANALYSIS Haldia, 3. Munger, 4. Jamalpur) under investigation (Standard False Colour Composite LISS-III images with Green, Red and 4.1 Responses of HH/HV polarizations NIR bands in Blue, Green and Red channels respectively). The correlation graph between the HHσ0 and HVσ0 in Figure 2 SAR data from Fine Beam Dual polarized HH/HV L-band shows that the higher values correspond to built-up class; while ALOS PALSAR imageries are used in this study for all the generally much higher for HH in comparison to HV. The least sites. Specifications of the data are mentioned in Table 1. values correspond to the water class, while vegetation and bare soil lie in between. Remote sensing data type SAR Satellite ALOS (a) (b) Sensor PALSAR Launching country Japan (JAXA) (Organization) Date of launch 24th January, 2006 Spatial resolution 25m Swath width 70km (34.3° incident angle) L-band, 15-30 cm; Fine Beam Wavelengths; polarization Dual (HH/HV) Year of data acquisition 2010 Source of data acquisition JAXA, Japan Figure 2. Relation between HHσ0 (x-axis) and HVσ0 (y-axis) showing built-up class encircled in ‘red’ over (a) Kolkata- Table 1. Satellite data specifications Haldia site and (b) Munger-Jamalpur twin city site. This contribution has been peer-reviewed. The double-blind peer-review was conducted on the basis of the full paper. https://doi.org/10.5194/isprs-annals-IV-5-455-2018 | © Authors 2018. CC BY 4.0 License. 456 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume IV-5, 2018 ISPRS TC V Mid-term Symposium “Geospatial Technology – Pixel to People”, 20–23 November 2018, Dehradun, India index values for the major LULC classes over all the study sites 4.2 Classification using SAR-based index are shown. It is evident from the figure that the values for built- up areas are non-overlapping with that of other classes, and is The proposed SAR derived index of Equation 2 is applied on purely negative unlike other classes. the study site PALSAR imageries and the output maps, after binary classification into built-up and non built-up areas are illustrated in Figure 3. Built-up areas are demarcated in ‘black’ and non built-up areas in ‘white’. Figure 4. Index values (y-axis) for feature classes. Built-up Vegetation Bare soil Water Mean -0.90003 0.142524 0.245075 0.237397 SD 1.118926 0.100006 0.05316 0.059535 Table 2. Mean and SD values of index for major LULC classes Figure 5. Index values (y-axis) for feature classes. 4.4 Classification accuracy Figure 3 demonstrates the classified binary image illustrating the extracted built-up area marked in ‘black’ and the non built- up area in ‘white’. Accuracy assessment for the classification is Figure 3. Classified binary images from the SAR-based performed. A total of 900 sample points are randomly selected indexed maps distinguishing built-up (in black) from non built- over the sites as reference points. Maximum likelihood up (white) areas. algorithm is used for supervised classification (Sinha et al., 2013) and the accuracy is recorded in terms of overall accuracy 4.3 Responses of the index (OA) and kappa coefficient (k) for classification of built-up and non built-up areas (Table 3). Overall accuracy of classification The responses of the index for the sites over different land use is 91.89% with k value of 0.83. land cover types are recorded and a graph is designed, as depicted in Figure 4. The red line in the graph shows the built- Built-up Non Built-up OA k up class that can be clearly distinguished from the other classes, viz. water, vegetation and bare soil. Built-ups are visualized as Built-up 301 41 342 91.89 0.83 a complete separate class with mostly negative values of the Non Built-up 32 526 558 index; while all the remaining classes show a positive index 333 567 900 value. Mean and Standard Deviation (SD) of the index values shows negative mean values only for built-up areas, while 900 positive for the rest classes (Table 2). This has also been represented in Figure 5, where maximum, minimum and mean Table 3. Accuracy assessment report This contribution has been peer-reviewed. The double-blind peer-review was conducted on the basis of the full paper. https://doi.org/10.5194/isprs-annals-IV-5-455-2018 | © Authors 2018. CC BY 4.0 License. 457 ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume IV-5, 2018 ISPRS TC V Mid-term Symposium “Geospatial Technology – Pixel to People”, 20–23 November 2018, Dehradun, India 5. CONCLUSIONS Corbane, C., Faure, J.-F., Baghdadi, N., Villeneuve, N., Petit, M., 2008. Rapid urban mapping using SAR/optical imagery Efficient delineation of built-up areas is of great ecological and synergy. Sensors, 8 (11), pp. 7125-7143. environmental importance but is a challenge that researchers doi:10.3390/s8117125. experience. Worldwide limited applicability of multi-spectral optical remote sensing for automatic extraction of built-up Hu, H., Ban, Y., 2008. Urban land-cover mapping and change features led into the intervention of SAR technology in this detection with Radarsat SAR data using neural network and research. The study illustrated an innovative approach in rule-based classifiers. In: XXI Congress of International Society effectively delineating built-up areas with SAR derived index for Photogrammetry and Remote Sensing (ISPRA). Beijing, using dual polarimetric properties of HH and HV polarizations China, pp. 1549-1553. of ALOS PALSAR data. The method is developed, tested and validated over multiple sites. It resulted in about 92% overall Liu, C., 2016. Analysis of Sentinel-1 SAR data for mapping accuracy in mapping built-up areas. The method adopted is standing water in the Twente region. M.Sc Thesis. University of acceptable, simple, reliable and replicable with high accuracy in Twente, Enschede, The Netherlands. mapping the built-up features. Lv, Q., Dou, Y., Niu, X., Xu, J., Xu, J., Xia, F., 2015. Urban The study is important for change detection of built-up areas, land use and land cover classification using remotely sensed urban management and policy making, LULC mapping, urban SAR data through deep belief networks. Journal of Sensors, sprawl assessment, urban microclimate studies, urban heat 2015(10). doi:10.1155/2015/538063. island, etc. and other related studies. Proper assessment of the spatial extent of built-up areas is essential for accurate Qin, Y., Xiao, X., Dong, J., Chen, B., Liu, F., Zhang, G., understanding and calculating these interconnected built-up Zhang, Y., Wang, J., Wu, X., 2017. Quantifying annual changes attributes. With the arrival of various SAR missions like ALOS in built-up area in complex urban-rural landscapes from and RISAT follow-ups, NISAR, MAPSAR, etc., which can analyses of PALSAR and Landsat images. ISPRS Journal of provide circular and quad polarizations along with single and Photogrammetry and Remote Sensing, 124, pp. 89-105. doi: dual polarization for further improvements in this research. 10.1016/j.isprsjprs.2016.12.011. ACKNOWLEDGEMENTS Shao, Z., Fu, H., Fu, P., Yin, L., 2016. Mapping urban impervious surface by fusing optical and SAR data at the We sincerely acknowledge Japan Aerospace Exploration decision level. Remote Sensing, 8 (11), pp. 945. Agency (JAXA, Japan) for providing the SAR data. We are also doi:10.3390/rs8110945. thankful to the Geoinformatics Cell of Department of Civil Engineering, Haldia Institute of Technology (India) where the Sinha, S., Jeganathan, C., Sharma, L.K., Nathawat, M.S., research was performed. We express sincere gratitude to 2015b. A review of radar remote sensing for biomass Science and Engineering Research Board (SERB), Department estimation. International Journal of Environmental Science and of Science and Technology (DST), Government of India for Technology, 12(5), pp. 1779-1792. doi:10.1007/s13762-015- providing funds under SERB National Post-Doctoral fellowship 0750-0. (SERB NPDF) scheme (File Number: PDF/2015/000043). Sinha, S., Jeganathan, C., Sharma, L.K., Nathawat, M.S., Das, REFERENCES A.K., Mohan, S., 2016. Developing synergy regression models with space-borne ALOS PALSAR and Landsat TM sensors for Abdikan, S., Sanli, F.B., Ustuner, M., Calò, F., 2016. Land retrieving tropical forest biomass. Journal of Earth System cover mapping using Sentinel-1 SAR data. In: The Science, 125(4), pp.725-735. doi:10.1007/s12040-016-0692-z. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. XLI-B7, pp. 757-761. Sinha, S., Santra, A., Mitra, S.S., 2018. Automated extraction of doi:10.5194/isprs-archives-XLI-B7-757-2016. built-up areas within forests using remote sensing. In: A. Santra, N.K. Yadav, eds. Proceedings of National Conference on Aghababaee, H., Niazmardi, S., Amini, J., 2013. Urban area Advancement in Civil Engineering Practice and Research. extraction in SAR data. In: The International Archives of the Excel India Publishers, New Delhi, India. pp. 96-99. Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. XL-1/W3, pp. 1-5. doi:10.5194/isprsarchives- Sinha, S., Sharma, L.K., Nathawat, M.S., 2013. Integrated XL-1-W3-1-2013. geospatial techniques for land-use/land-cover and forest mapping of deciduous Munger forests (India). Universal Bramhe, V.S., Ghosh, S.K., Garg, P.K., 2018. Extraction of Journal of Environmental Research & Technology, 3, pp. 190- built-up areas using convolutional neural networks and transfer 198. learning from Sentinel-2 satellite images. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Sinha, S., Sharma, L.K., Nathawat, M.S., 2015a. Improved Information Sciences, Vol. XLII-3, pp. 79-85. land-use/land-cover classification of semi-arid deciduous forest doi:10.5194/isprs-archives-XLII-3-79-2018. landscape using thermal remote sensing. The Egyptian Journal of Remote Sensing and Space Science, 18(2), pp. 217-233. Chen, Z., Zhang, Y., Guindon, B., Esch, T., Roth, A., Shang, J., doi:10.1016/j.ejrs.2015.09.005. 2013. Urban land use mapping using high resolution SAR data based on density analysis and contextual information. Wang, A., Liu, P., Xie, C. 2016. Urban land use classification Canadian Journal of Remote Sensing, 38(6), pp. 738-749. from high-resolution SAR images based on multi-scale Markov doi:10.5589/m13-002. Random Field. In: 24th International Conference on Geoinformatics. Galway, pp. 1-4. This contribution has been peer-reviewed. The double-blind peer-review was conducted on the basis of the full paper. https://doi.org/10.5194/isprs-annals-IV-5-455-2018 | © Authors 2018. CC BY 4.0 License. 458