Classification for polarimetric SAR images based on subaperture decomposition

In this paper, a novel method, which combines subaperture decomposition with H/α/Wishart classifier, is introduced to classify polarimetric synthetic radar (SAR) images. We use H/α plane to initially classify the full-resolution polarimetric SAR image. The initial classification map defines training sets for classification based on coherency matrices of subapertures and wishart distribution. The classified results are then used to define training sets for the next iteration. The convergence of this kind of iteration is much faster than using full-resolution images. And the interpretation of this method is more easily. Results show that the method provides a significant performance improvement with respect to the approaches based on full-resolution images.