Polarimetric SAR images clustering with Gaussian mixtures model

摘要

Polarimetric Synthetic Aperture Radar (PolSAR) images offer a possibility to detect the right target backscattering giving a better classification results. In this paper we are interested in polarimetric SAR image clustering by using model-based polarimetric decomposition of Freeman and Yamaguchi like feature vector with the Gaussian Mixture Model (GMM) unsupervised classifier. For the validation of our results we use the polarimetric SAR images of the oberpfaffenhofen site. The results obtained from real data are significant.