Applying the log-cumulants of texture parameter to fully polarimetric SAR classification using Support Vector Machines Classifier

In this paper, we proposed a fully polarimetric SAR classification method based on the log-cumulants of texture parameter of the fully polarimetric SAR data. Unlike other classification algorithms that classify pixels by their scattering characteristics, this method will use a combination of the texture parameter of fully polarimetric SAR data and the Support Vector Machines (SVM) Classifier based on the spherically invariant random vectors (SIRV) model. A full polarimetric image Oberpfaffenhofen region in Germany, acquired by E-SAR at L-band, is used for our experiment. It is shown that the proposed method is consistent with the actual scattering mechanisms, especially for urban areas, and can be used to effectively distinguish different types of terrains.