An efficient method of classification of fully polarimetric sar images

Purpose. Synthetic Aperture Radar (SAR) can be used to acquire high-resolution images of ground targets during night as well as day or in good weather as well as inclement weather, so it plays an important role in the national economy and defense. However, according to the characteristics of SAR imaging mechanisms, a geometric distortion and a form of multiplicative noise, known as coherent speckle, generally corrupt the resulting image. SAR image classification is the foundation of SAR image interpretation. For the existing of speckle, traditional image classification technologies cannot work well. In this paper, in order to further improve the effect of polarimetric SAR image classification, an efficient classification method of fully polarimetric SAR image that is based on polarimetric features, the scattering intensity information, and Fuzzy C-Means (FCM) Algorithm is proposed. Methodology. Combining the scattering properties of fully polarimetric SAR image with the scattering intensity information, the total scattering power, based on H/α/A/SPAN(H, Entropy; α, Scattering angle; A, Anisotropy degree; SPAN, the total power of polarization), we obtained the initial classification result of polarimetric SAR image. Then with FCM Algorithm, the result of the polarimetric SAR image classification was achieved. Findings. The experimental results show that the proposed method is superior to the traditional methods of fully polarimetric SAR images classification. Originality. The proposed method not only considers the scattering properties of fully polarimetric SAR data but also combines the statistical characteristics information. The proposed method provides good result of classification of polarimetric SAR image and, to some extent, keeps the scattering properties. Practical value. The experiments have proved that the proposed algorithm can keep the texture and details of SAR image better, can give better classification result to the traditional classification methods of fully polarimetric SAR image. The proposed method is useful in SAR image interpretation. © 2015 Xiaorong Xue, Wei Wang, Hongfu Wang, Fang Xiang.