由于合成孔径雷达系统的有限带宽，使阶跃边缘变成斜坡型边缘，线条型边缘变成屋顶边缘，斜坡型边缘甚至难以探测。在本文中针对合成孔径雷达图像，提出了一种新型的边缘提取方法。这个方法分为二个步骤；首先根据图像亮度进行局部分割（如在一个5*5的窗口内），得到亮度不同的两个区域。第二步：将亮度不贩 两个区域边界作为图像的边缘，并根据亮度变化确定图像的边缘，将这个方法用于合成孔径雷达图像边缘提取，成功地提取了合成孔径雷达图像的边缘信息。

Since SAR sensors have limited bandwidths, this leads to slow responses to sudden changes ( smearing sharp edges) . Thus, it is difficult to detect edges in SAR images. A new edge detection method is developed to detect edges in SAR images in this paper. There are two steps to accomplish it. First, the local area is segmented into one bright area and one dark area. Second, the boundary of the bright area and the dark area is regarded as the edge, and the edge is identified based on intensity variety. Application of this method to SAR images has shown that the method can effectively detect edges.