Contour lines recognition from scanned topographic maps
Runsheng Wang · Computer-aided engineering · 2005
Aliasing and false colors in the scan of a common-conditioned map are so serious that when contour lines is segmented from it, not only many gaps due to contour lines intersecting and/or overlapping with other features become more intricate, but also thick lines (lines conglutination) due to contour lines closely spacing occur. An effective method is presented for contour lines recognition from scanned common-conditioned topographic maps. It consists of three distinguished aspects. Firstly, a color key set technique is adopted in color map segmentation to compensate for aliasing and false colors. Secondly, local window segmentations are performed at the place where the gaps and thick lines occur in the original image to resolve contour lines disconnection and conglutination. Finally, the global topologic properties of contour lines are used to filter the accurate contour lines. The experimental results show that this new method is potential when applied to a number of well-conditioned and common-conditioned topographic maps.