合成孔径雷达图像统计滤波降噪方法

A Speckle Reduction Technique of SAR Images Based on Statistic

基于统计滤波方法需要在尽可能大的同质区域内进行滤波处理的要求 ,根据 MHR和 EPOS算法采用的可变窗口思想 ,提出了从八个不同方向上逐步扩大窗口 ,直到出现边界的搜索算法 ,在保持图像边界信息的同时滤除了斑点噪声的影响 .对于噪声统计量 ,提出了利用了比例图来估计其大小的方法 ,并由此求出搜索的域值 .实验结果表明 ,该方法处理的

Speckle is the dominant effective factor to the quality of SAR images. The use of statistical property of image proposes an efficient method in spatial domain for speckle filtering and image restoration. Because the effects of both speckle filtering and edge preserving are interactional, the purpose of current filters is to improve efficiently such two factors. The classical filtering method such as kuan and Gamma MAP work efficiently to reduce the noise, but at the same time they smooth, to some extent, the edge information. Both MHR and EPOS method devise different adaptive windows to make up for this deficiency. Through the intensive research on the classical adaptive filter, this paper proposes an integrated technique that is applied on reduction of speckle and restoration of image efficiently based on adaptive window. The result of experiment indicates that the proposed method improves the quality of SAR images with conspicuous effect.