Classification of Polarimetric Radar Data with Texture Measure

In this paper we have classified polarimetric SAR image utilizing it capability to capture image in different polarization. Different land covers exhibit different sensitivity to differently polarized wave. Therefore, statistical measures (mean, variance and contrast) have been used to identify the best polarization to classify a particular land cover class (urban, water, vegetation, and canal). Statistical measure were applied on different polarized image choosing window size of 7 X 7. Based on texture measure, a decision tree was formed to classify the polarimetric SAR image. Good overall and class accuracies were observed for classified image.