基于ROEWA与Hough变换的SAR图像边缘检测

Edge detection in SAR images based on ROEWA and Hough transform

指数加权平均比率(ROEWA)算子是SAR图像常用多边缘检测算子之一,但该算子只能检测图像的边缘强度,无法得到边缘的方向。为此利用Hough变换优秀的抗噪性能和方向选择性,提出了一种基于Hough变换的SAR图像边缘方向计算方法。首先使用ROEWA算子对SAR图像进行边缘检测,然后在传统非极值抑制定位方法的基础上,进一步细化边缘,最后利用Hough变换对细化后图像计算边缘方向。SAR图像对比实验结果表明,该方法能够更准确地计算边缘位置与边缘方向,为SAR图像线性目标检测提供了一条有效途径。

Ratio of exponential weighted average（ROEWA） operator is one of multi-edge detector and is widely used in SAR images.But the detector can only detect the edge strength,cannot get the edge direction.So using the excellent noise immunity and directional selectivity of Hough transform,a calculation method of edge direction based on Hough transform is proposed in this paper.Firstly,the edge result with ROEWA operator from SAR image is obtained,and the edge result on the basis of traditional non-maxima suppression location algorithm is thinned.Finally,Hough transform is used to compute the edge direction about the thinned image.Comparison of experimental results on SAR images proves that the method can calculate the edge location and direction more accurately,and provide an effective way to detect linear target in SAR images.