本文阐述了一种基于二进小波变换的从SAR图像中提取湖岸线的方法。该方法首先利用边缘和噪声的奇异性在二进小波变换中随尺度变化具有不同的传播规律来检测边缘点，然后利用基于梯度矢量流的主动轮廓模型进行边缘点的连接，实验结果证明，这种方法可以有效的消除SAR图像中存在的斑点噪声对边缘提取的影响，准确的提取湖岸线。

In the field of map production, flood surveillance, and change study of lakeshore, the lake shoreline must be detected. A new method is explored in this paper, which can detect the lake shoreline in SAR images. Using this method, The edge points are detected by Mallat' s Wavelet-based edge detection method firstly, and then the Gradient Vector Flow Active Contour Model is used to connect the edge points. The result of the experiment shows that method presented in this paper can restrain the interference of speckle noise, and detect the lake shoreline precisely.