高分辨率SAR影像Frost时相滤波方法

Frost Multi-temporal Filtering Method for High-resolution SAR Imagery

高分辨率SAR影像中大量存在的相干斑噪声是制约图像解译、点目标识别和差分干涉测量精度的重要因素之一。为了有效地抑制和去除SAR影像中的斑点噪声,基于以往研究,采用了一种基于Frost滤波定权的多时相滤波方法(即Frost时相滤波)。该方法利用GAMMA软件配准多时相SAR影像,然后利用Frost滤波为Multi-Temporal滤波方法定权,最终获得Frost时相滤波后的影像。实验选用分辨率为1.3 m×1.8 m和多视比为1∶1的TerraSAR-X卫星影像,选取乡村田地和城镇作为研究区域,获取传统Frost滤波、GAMMA多时相滤波和Frost多时相滤波算法的高分SAR影像滤波结果。在GAMMA和Matlab软件中进行对比分析实验,通过目视解译,以及平滑指数、相干斑指数、等效视数、归一化指数、辐射分辨率等量化评价指标分析,验证了Frost时相滤波方法能有效抑制高分SAR影像中的斑点噪声,并明显改善高分SAR影像质量。

The speckle noise in high resolution SAR images is one of the important factors that restrict the accuracy of image interpretation,point target recognition and differential interferometry.In order to effectively suppress and remove speckle noise in SAR images,based on previous studies,a weighted multi-temporal filtering method based Frost filter(i.e.,Frost temporal filtering)has been adopted.First,the multi-temporal SAR images are registered by GAMMA software in this method,then Frost filter is used to set the weight for the multi-temporal filtering method,and finally the images after Frost Temporal filtering are obtained.In the experiment,Terrasar-x satellite images with a resolution of 1.3 m×1.8 m and a multi-look ratio of 1∶1 are selected.Rural fields and urban areas are selected as the study areas,obtaining the high-resolution SAR image filtering results of traditional Frost filtering,GAMMA multi-temporal filtering and Frost multi-temporal filtering algorithms.The experiment is processed in GAMMA and Matlab software,the experiment results are compared and analyzed through visual interpretation and quantitative evaluation index analysis such as smoothness index,speckle index,equivalent number of looks,normalized index and radiation resolution.The Frost temporal filtering method can effectively restrain the speckle noise in SAR image,and the SAR image quality is obviously improved.