SAR image change detection based on Brushlet transform

The traditional change detection method has a poor accuracy for similarity character capture and low direction-resolution. In this paper, a new 2D-Otsu SAR image change detection method is proposed based on the overcomplete Brushlet transform and Gabor window. This method combines the local anisotropic Gabor weighted nonlinear mean procedure in the overcomplete Brushlet domain and linear combination with the minimum mean squared error in the original domain to obtain mean character after the speckle noise is removed, which resolves the problem of low direction-resolution, and can accurately position the texture of each direction, frequency and position. Finally, change detection is processed by the 2D-Otsu method which combines the mean character and gray-level character. Experiment results show that the new method has a better performance, and can well preserve the detailed information such as the texture and edge.