SAR Image Thresholding Based on 2-D Fuzzy Tsallis Entropy and Chaotic Particle Swarm Optimization

摘要

Aiming at the problem that thresholding based on onedimensional(l-D) histogram is sensitive to noise and can't reflect spatial information of synthetic aperture radar(SAR) images,a maximum fuzzy Tsallis entropy thresholding method based on two-dimensional (2-D) histogram is proposed. This method considers the spatial information of SAR images and new membership functions are applied. Furthermore, the formula for maximum fuzzy Tsallis entropy thresholding based on 2-D histogram is derived. In view of the shortage that the amount of computation of 2-D maximum fuzzy Tsallis entropy thresholding is large, the optimal thresholds are obtained by chaotic particle swarm optimization (CPSO). The experimental results show that this method has advantages in respect of segmentation results and running time.