Algorithms for automatic vectorization of scanned maps

The proliferation of Geographic, Information System (GIS) in industry and research has lead to the need for converting the available analog geospatial data to digital form. Although maps can be scanned, they cannot be used directly in a GIS system without processing. All the available commercial raster to vector conversion software are semi-automatic and require an operator for digitization and verification. Thus automatic algorithms are required for faster and reliable conversion of maps where the operator is required only for verification.; Methods for automatic vectorization of scanned maps deal with polygons, lines and points. The focus of this research is on the extraction of linear features from scanned maps and satellite imageries using skeletonization. Two methods that are examined are skeletonization by Voronoi Diagram and Mathematical Morphology. Connectivity is established between disconnected lines using Least Square Parabola (LSP). The objectives of this research are to compare the two vectorization methods and examine the application of LSP for gap filling. (Abstract shortened by UMI.)...