A Modified Change Detection Method Focusing on De-Speckling of SAR Images

SAR images are important data for observing the earth in the all weather conditions. But the speckle noise contaminated the SAR images obstructs the application of them. Speckle noise is primarily due to the phase fluctuations of the electromagnetic return signals. In this paper, we investigate the effect of speckle on the change detection result. Two cost functions are used in our change detection algorithm. One is based on Minimum Mean Squared Errors and the other one is composed of the MMSE and a de-speckled filter. Change detection method is based on the genetic algorithm. The difference image is first computed by the input images and then divided into two separated changed and unchanged regions, according to the change mask realization produced by Genetic Algorithm. A set of real data acquired from ASAR sensor of Envisat satellite is used to evaluate the performance our method. The results show that the proposed method improves the detection procedure reducing the noise contamination.