Unsupervised classification of PolInSAR image based on Shannon Entropy Characterization

In this paper, we propose a new method for unsupervised classification of polarimetric synthetic aperture radar interferometry (PolInSAR) images based on Shannon Entropy Characterization. Firstly, we use polarimetric H (entropy) and a parameters to classify the image initially. Then, we reclassify the image according to the span of Shannon Entropy Characterization. Finally, we fuse the results of the two previous steps and merge them to the specified number of clusters. The effectiveness of this method is demonstrated on CETC38 PolInSAR data and E-SAR PolInSAR data.