Classification based on four-component decomposition and SVM for PolSAR images

A new algorithm of target classification for polarimetric SAR data is proposed in this letter. First, each pixel is decomposed into four scattering components which are used for the feature vectors. Second, classifier can be designed using support vector machines through training the selected samples and then applied in segmentation of the images to be tested. The experiments are used for analysis, which are carried out on polarimetric data from the NASA/JPL AIRSAR of San Francisco.The results indicate it is feasible and efficient that combining four-component decomposition and SVM for PolSAR image classification.