Polarimetric SAR Data Classification with Eigenvalue-based Decompositions
基于特征值分解的全极化SAR遥感数据分类研究

This paper applied the eigenvector\|based decompositions method,and studied the fine classification problem of the polarimetric SAR imagery.Under the analysis of H-Alpha plane,we introduced another characteristic parameter named the mean target power (span) which was presented with all of eigenvalues and pseudo\|probabilities.And a simple method of selecting thresholds to segment the mean target power of the targets was given,and the Radarsat-2 data of Poyang Lake area was used to test the method.After the analysis of classification results,we found that the method firstly considering the mean power information cound provide much more detailed classification than H-Alpha plane classification with only scattering entropy and angle information.