URBAN IMPERVIOUS SURFACE DETECTION WITH LANDSAT AND FULL POLARIMETRIC ALOS/PALSAR IMAGES
Keywords: Urban impervious surfaces, Full polarimetric SAR, Landsat, Texture, Neural network, Classification
Abstract. Impervious surface detection in urban area is of great importance because impervious surfaces have been recognized as an urbanization index. In other words, an increase of impervious surface extent accounts for urban sprawl. Remote sensing images in optical wavelength proved to be beneficial for urban impervious surfaces detection and mapping. However, the use of optical images is limited in cloud-prone regions. Also, there are some complexities which make it challenging to impervious surface detection especially in urban areas. In recent years, in the light of increasing availability of SAR data, the use of SAR data for urban impervious surface detection is increased as well. In this study the efficiency of full polarimetric L-band images was compared to Landsat images. Furthermore, the role of SAR’ texture features was assessed. Results show that the classification accuracy of SAR images cannot overcome the Landsat’s classification accuracy. However, SAR texture features provide positive contribution in impervious surface detection in homogeneous urban region.