SAR Change Detection Based on Cluster Distribution Divergence
-
Graphical Abstract
-
Abstract
A new method to detect the change of two SAR images using cross entropy based on the statistical character of SAR cluster is proposed.Rayleight distribution model is used to describe the SAR cluster,and the different index of two images is derived using Kullback-Leibler divergence in detail.Constant false alarm rate is also introduced to segment the change areas.The method is demonstrated feasible using the RadarSAT SAR image.
-
-