A Watershed-based Automatic Building Detection in High-resolution Airborne Urban SAR Images
基于分水岭变换的高分辨率机载城区SAR图像建筑物自动检测方法

The automatic detection of building boundaries in a single high-resolution urban SAR (Synthetic Aperture Radar) image based on watershed transform with markers is proposed in this paper. According to the strong radar backscatter energy and shape feature of a building in SAR images, markers are extracted mainly by a CFAR detector and direction-correlation analysis. Then the minima imposition is used to modify the gradient of the initial SAR image with the help of markers. Finally, the building boundaries are obtained by computing the watersheds of the modified gradient image. The experiment shows the feasibility of our method.