Integration of polarimetric decomposition, object-oriented image analysis, and decision tree algorithms for land-use and land-cover classification using RADARSAT-2 polarimetric SAR data

A novel method which integrates polarimetric decomposition, object-oriented image analysis, and decision tree algorithms is presented for land-use and land-cover (LULC) classification using RADARSAT-2 polarimetric SAR (POLSAR) data. Polarimetric decomposition which is aimed at extracting polarimetric parameters related to the physical scattering mechanisms of the observed objects can be used to support the classification of POLSAR data. The main purposes of object-oriented image analysis are delineating image objects as well as extracting various textural and spatial features from image objects to improve classification accuracy. A decision tree algorithm provides an efficient way to select features and implement classification. Compared with the Wishart supervised classification which is based on the coherency matrix, the proposed method can significantly improve the overall accuracy and kappa value of LULC classification by 17.45 percent and 0.24, respectively. Further investigation was carried out on the contribution of polarimetric decomposition, object-oriented image analysis, and decision tree algorithms to the improvement achieved by the proposed method. The investigation shows that all these three methods contribute to the improvement achieved by the proposed method.