Abstract:For improving the estimated precision and rationality of the scattering coherency matrixes of the class center of different terrain types among the unsupervised classification based on polarimetric target decomposition and the maximum likelihood classifier based on the complex Wishart distribution, a new method for unsupervised classification of terrain types is proposed in this paper. Through weighted-based calculating coherency matrix of the center of every class, the coherency matrix can better represent the class center because the method well considers the correlation of pixels and the texture information of SAR image. The algorithm is described in detail and the contrastive experiment is done using AIRSAR L band polarimetric images. The experiment result indicates that the classification’s accuracy of the method is higher and the iterative speed is faster.