A study on extraction of urban areas from polarimetric Synthetic Aperture Radar image

This paper discusses the polarimetric correlation coefficient to extract the urban areas from polarimetric Synthetic Aperture Radar (POLSAR) image. For classification of POLSAR image, several methods have been proposed to extract polarimetric feature, such as Polarimetric Entropy-Alpha, three-component scattering model, Huynen parameters and so on. However, there is a possibility that the polarimetric correlation coefficient has a potential for the objective of this paper, too. In order to verify the capability of polarimetric correlation coefficient, we examine the behavior of this coefficient between the urban areas and the natural distributed areas with respect to the several polarimetric scattering models and the difference of polarization basis. Moreover, we apply the polarimetric correlation coefficient to the actual polarimetric SAR data acquired by Pi-SAR/X-SAR.