The various applications have been promised by the new generation of the spacecraft SAR data (i.e. Radarsat2 and TerraSAR-X), as the classification, the decomposition, and the modeling of the polarimetric synthetic aperture radar (SAR) data has been improved in recent years .This work is based on the fact that in order to extract the various patterns in distinctive field of studies, all the scattering matrix components are informative source of data. Different cross products of the complex scattering matrix channels (HH, HV, VH, and VV) that are involved in the phase and amplitude information are joined together to build instructive features. In the vector space, Fisher class separability algorithm will be tested, and the features with the best class separability, large distance between classes, and small within-class variances will be selected. As we measured the classification effectiveness of the individual features, we needed to choose a subset of the informative features from the nine originally available features. Finally, we combined all the educational information contents in order to classify desired images with the best overall accuracy.
