Small-scale Cartographic Generalization Research Of Road Based On Constraint

Along with the rapid developments of computer science and technologyand their increasingly intensive applications in cartography, the theories andmethods of cartography are also constantly enriched. In the digital technologyenvironment, cartographic generalization is gradually developing towards tothe direction of automation. Compared with the traditional cartography,automatic cartographic generalization has great prospects of development.Meanwhile it is also facing a number of challenges and difficulties. Theautomatic integration of liner elements has been always the emphasis anddifficult points in the research field of cartographic generalization. Road is themain linear elements, so the research on road cartographic generalization hasvery important meanings.Base on a comprehensive study of the theories and methods of automaticcartographic generalization, this thesis made some research and explorationson the constraint-based small-scale road generalization, and established atopological similarity model for roads selection, then chose the appropriatesimplification approach to do experiments. The main researching contents andachievements are as follows:First, this thesis described the constraint-based automatic cartographicgeneralization theories, then discussed the constraints required to follow:classification and grading constraints, geometric constraints, structuralconstraints, topological constraints and Gestalt constraints.Second, based on the comprehensive geometry and structural constraints,this thesis identified the bending for the roads before the simplificationgeneralization to full possibly retained the road curved features. Therefore,this thesis elaborated three curved recognition methods: bending recognitionbased on inter-visibility, convex curved recognition and bending recognitionbased on maxima curvature. And advantages and disadvantages of these threemethods were compared.Third, Douglas-Peucker algorithm was improved on the aspect of thethreshold selection, and this were proved reasonable and feasible by theexperimental results. Meanwhile, the methods after improvements got betterroad generalization quality and enhanced the computing speed.Fourth, the constraint principles of road generalization were quantified,and quality evaluation parameters were provided: geometric accuracy evaluation and position accuracy evaluation to evaluate the results ofcartographic generalization.Experimental results show that the design and implementation of this system onroad cartographic generalization were reasonable on the selection of operators andalgorithms, and more importantly, they can effectively improve the overall efficiencyand overall quality of road mapping.