Automatic interpretation of scanned maps: Reconstruction of contour lines

This paper deals with the problem of automatic reconstruction of contour lines out of scanned maps. Our approach is essentially based on the global topology of the contour lines (i.e. a set of nonintersecting closed lines), but we use also some local and geometrical informations to guide the reasoning. Our proposed method starts from a map scanned in RGB. The image is first segmented by a colour classification process, and the objects are skeletonized end vectorized. Then, by a local analysis, we weight and add some hypothetical edges. By a relaxation process, we try to orientate the elevation curves, and so we could re-adjust the weight. Finally, we found a topologycally correct solution by resolving the dual problem (i.e. a bi-colouring problem).