基于模型与算法的结构化地图信息自动综合研究

The main objectives are to study the basic models and algorithms of geo-information generalization, to acquire the spatial knowledge and its formalization and application in generalization decision making. The main research results are as follows: establishing the fundamental models of generalization which were implemented in form of three sub-models of generalization: creating the global structure to guarantee the correct selection of geographic objects; creating the local structure to guarantee the weight evaluation for every individual objects; portraying the object through discovering the inner micro structure to carry out the object generalization. The implementation algorithms are represented in forms as follows: construction the embedded convex polygons for point objects for representing the spatial distribution characteristics; tree structure for line network map features; <attraction principle> for amalgamation of area objects and contour tree structure for representing the relief structure. Using the Voronoi diagram the difference among the objects belonging to the same tree level can be distinguished. All these measures are to represent the global structuring in all aspects. At last the application of fractal method to deduct the variation law with the map scale change , the application of knowledge acquisition, its formal description and spatial reasoning in map design and in decision making of map generalization. The main results will be published in title <Fundamental Models and Algorithms of Generalization for GIS and Map Information >.

建立基于模型与算法的结构化制图综合自动化的理论与技术方法体系，从理论基础上研究GIS 的多比例表达与多尺度数据库问题，以支持多层次规划决策，克服GIS 进一步发展的障碍。要解决的基本问题是：利用单回归与复回归，解决数量综合问题，利用当代高新技术，如图论拓扑、计算几何与计算机视觉，解决综合选取的质量（结构与地理分布）问题。