Road extraction in high resolution remote sensing images based on mathematic morphology and snake model

This paper proposes an algorithm for road extraction in high resolution remote sensing images based on mathematic morphology and snake model, which can divide up the area of objective directly from an image and combine with snake model to extract the road. Accordingly, the first step consists in extracting the road area from the high resolution remote sensing image based on morphology. In a second step, get the contour line of the area as the initial contour of Snake to track the road. Experiments turn out that the algorithm is extremely fast in image division and effective in extracting roads in high resolution remote sensing images.