The discriminatory capability of polarimetric SAR data for land use classification

A SIR-C data is used to assess the discriminatory capability of full polarimetric data for several classes of land-use. It is analysed the contribution of each type of data (phase difference, intensity ratio, intensity pair and intensity-phase difference pair), using the iterated conditional modes (ICM) classifier. It is shown that each class was better classified using a different type of polarimetric data. The result of the classification (measured by the confusion matrix and the Kappa coefficient of agreement) was considered very good, allowing the discrimination of nine land use classes, which includes different cultivation stages of some crops