改进的基于结构检测的SAR图像去斑方法

An Improved Method of Speckle Filtering in SAR Image Based on Structure Detection

合成孔径雷达图像固有的相干斑噪声严重降低了图像的可解译程度,影响了后续目标检测、分类和识别等应用。因此,SAR图像的相干斑抑制问题一直是SAR图像应用的重要课题之一。一个理想的去斑算法应该在平滑的同时保持图像的边缘等细节不受损失,目前存在各种各样的算法,但没有一种方法能够完美的满足这一要求。为此该文提出了一种改进的结构检测的SAR图像去斑算法。利用概率迭代方法分割图像并检测边缘,结合强点检测图,将SAR图像标为结构区和非结构区,在非结构区域内进行Lee滤波以平滑噪声,对结构区直接保留原值,获得了非常好的去斑效果。利用RADARSAT实测图像进行实验,并对实验结果作充分分析,证明了本算法的有效性。

Speckle, appearing in synthetic aperture radar （SAR） images as granular noise, is generated by the coherent processing of radar signals. It severely affects the image qualities. To alleviate deleterious effects of speckle, various ways have been devised to suppression it. An ideal algorithm should smooth the speckle without blurring edges and fine detail, But most algorithms cannot meet these two requivements very well. This paper presents an improved method of speckle filtering in SAR image based on structure detection to solve some problems of the Lopes＇ structure detection and statistical adaptive speckle filtering, The method uses the probability iterative to segment SAR image and detect the edges, then combines the strong scatterer detection result to label the SAR image as structure area and non - structure area. In non - structure area, the lee filter is used, and in the structure area, the intensity of the image is preserved. Experimental results with RADARSAT images verify the practicability and the advantage of the improved method.