基于改进Frost滤波的SAR图像斑噪抑制算法

SAR Image Noise Suppression Based on Improved Frost Filtering Algorithm

针对合成孔径雷达(SAR)图像斑点噪声严重影响目标检测、识别、变化检测等有用信息的提取这一问题,通过对Frost滤波算法进行深入分析,提出了改进的Frost滤波算法。该算法的改进主要对滤波器的权值作了两方面的调整:一是引入像元距离贡献度参数,将此参数的倒数与滤波窗口内对应的像元与中心像元之间的欧氏距离作积融入滤波器权值的计算;二是将滤波窗口内的像元灰度统计特性考虑入内,将像元与中心像元的灰度差融入滤波器权值的计算。改进后的滤波器权值可以自适应地调整,避免了原始Frost算法的滤波盲目性,在斑噪抑制能力提高的同时也改善了边缘保持的效果。仿真实验结果表明,所提改进Forst算法在斑噪抑制和边缘保持方面均优于原始Forst算法。

In view of the problem of the synthetic aperture radar(SAR)image speckle noise impact SAR image target detection and recognition,change detection,by analyzing the Frost filter,the improved Frost filter algorithm is put forward.The improvement of the algorithm mainly done to filter a weight adjustment of two aspects:one is the introduction of contribution,like yuan distance to the bottom and filtering window like yuan between the centre and the corresponding pixels,Euclidean distance for the calculation of product into filter weights;the other one is to like yuan gray statistical characteristic of filtering window into filter weight calculation.The improved filter weights can be adaptive adjustment,avoid the blindness of filtering,original Frost algorithm in spot noise suppression capability as well as to improve the edge effect.Experimental results show that the proposed improved Forst algorithm on the edge of the spot noise suppression and is superior to keep original Forst filtering algorithm.