Urban road network extraction based on multi-resolution template matching and double-snake model

Most of the methods used so far for road extraction are to extract the roads as poly-lines from remote sensing image, but in the high resolution image, the road always has the character of surface, so the extraction result should be parallel contour. In the paper, multi-resolution analysis, template matching theory, combining double-Snake model are used to extract the road from high resolution image of QUICKBIRD. Firstly centralized poly-lines are extracted by template matching, and then the parallel contours are extracted by double-snake model and the road intersections are extracted. At last the whole road network is reconstructed. Experiment shows that the method is effective and practical.