Road and Linear Structure Automatic Extraction from Remote Sensing Images Using Marked Point Process

Automatic extraction of road and linear structure from remote sensing images is a very important problem. This paper uses a stochastic geometric model for automatic extraction of line network (roads, rivers...) from remotely sensing images. The line network in the observed scene is modeled by marked point process. Firstly, the thick branches are detected using a segment process. Secondly, a polygon tree is derived from this first result to represent the main part of the line network. Finally, new branch lines are extracted using a recursive algorithm based on the modeling of the descendants of a given branch line by a polyline process in the neighborhood of this branch line. Experimental results show the relevance of the object process models.