We propose an algorithm for estimating urban density from polarimetric synthetic aperture radar (SAR) images, and compare the urban density patterns of global megacities. SAR images are uniquely able to detect structural information of objects, but they are very sensitive to orientation angle. This issue has been an obstacle to applying SAR images to urban areas. Kajimoto and Susaki (2013b) proposed an algorithm to handle this issue. The effects of polarization orientation angle (POA) are removed by rotating the coherency matrix and then calculating the mean and standard deviation of scattering power by POA domain. The algorithm can estimate urban density from a single fully polarimetric SAR image but has the drawback that the generated urban density maps of multiple images are not comparable with each other because the algorithm generates a relative urban density valid only within the analyzed image. We therefore extend the method by calculating POA-domain statistics from all images of interest so that the generated maps can be compared. Estimated urban densities are assessed on two types of urban density generated from GIS data, building-to-land ratio and floor-area ratio. We demonstrate that the extended method can estimate urban density with reasonable accuracy. Finally, we generate two scattergrams of indices derived from urban density maps of global megacities. An analysis using the scattergrams indicates insightful information about the patterns of urban development. We conclude that the proposed algorithm and the analysis using the obtained results are beneficial to understanding the conditions in megacities.

No abstract

Commission I, III, VII WG I/2, ICWG III/VII KEY WORDS: PiSAR-2, Urban area, VHR SAR, Fully polarimetric SAR ABSTRACT:In this paper, we present a method to extract urban areas from X-band fully polarimetric synthetic aperture radar (SAR) data. It is known that very high resolution (VHR) SAR can extract buildings, but it requires more processes to map urban areas that should include other objects. The proposed method is mainly composed of two classifications. One classification uses total power of scattering and volume scattering derived by using four component decomposition method with correction of the polarization orientation angle (POA) effect. The other classification uses polarimetric coherency between SHH and SV V . The two results are intersected and final urban areas are extracted after post-classification processing. We applied the proposed method to airborne X-band fully polarimetric SAR data of Polarimetric and Interferometric Airborne Synthetic Aperture Radar System (Pi-SAR2), developed by the National Institute of Information and Communications Technology (NICT), Japan. The validation show that the results of the proposed method were acceptable, with an overall accuracy of approximately 80 to 90% at 100-m spatial scale.

In this paper, we compare the performance of extracting urban areas from L-band and X-band fully polarimetric synthetic aperture radar (SAR) images. For L-band SAR images, we used the method proposed by Kajimoto and Susaki [1] that combines the relation between volume scattering and total power and the polarization orientation angle (POA) randomness. For X-band SAR images, we compared two methods: the method proposed by Susaki and Kishimoto [2] and the method proposed by Esch et al. [3]. The former one replaces POA randomness used in the method by [1] with polarimetric coherence. The latter one calculates a texture index representing heterogeneity of land covers from single polarimetric SAR image. The validation results show that the accuracy for extracting urban areas depends on the heterogeneity of urban and vegetation areas.