SAR image segmentation based on Gabor filters of adaptive window in overcomplete brushlet domain

In this paper, a new technique based on Gabor filters with adaptive window is proposed for SAR image segmentation in overcomplete brushlet domain. SAR image is full of texture and direction information, and brushlet is a new kind of analysis tool for image with rich directional information. Aim at these characteristics, this paper combines Gabor filters with GLCP for segmentation in brushlet coefficients domain. In SAR image processing, the difficulty is the balance between regions and boundaries. The adaptive window technique is adapted in this paper to obtain the optimal size of Gabor filters. The presented method was found to be superior to three other methods by the experimental results for SAR image.