SAR图像中目标和背景的分割是SAR图像分割中的重要内容,阈值方法是其中比较简单实用的方法。阈值分割的核心问题是如何选择合适的阈值,最简单和常用的方法是从图像的灰度直方图出发,得到各个灰度级的概率分布密度,依据某一准则选取一个或多个合适的阈值,以确定每个像素点的归属。该文借助遗传算法工具,对基于二维直方图的模糊熵法做了改进和设计,提出了寻找最优阈值的分割算法,经MSTAR数据测试,对于含噪SAR图像目标和背景的分割具有很好的效果,抑噪功能强。 SAR image segmentation target and background is an important part of SAR image segmentation, threshold method is one of the more simple and practical method. The key problem of threshold segmentation is how to choose a suitable threshold. The simplest and most commonly used method is to start from the histogram of image grayscale, get the probability distribution density of each gray level, select one or more appropriate thresholds according to some criteria , To determine the ownership of each pixel. In this paper, genetic algorithm tools are used to improve and design the fuzzy entropy method based on two-dimensional histogram. A segmentation algorithm to find the optimal threshold is proposed. After MSTAR data testing, the segmentation of target and background with noisy SAR images is very important Good effect, noise suppression function.