Y. Long

School of Geographic Science, Nanjing Normal University, Nanjing  210097, China


Geographic information generalization is a necessary means for people to express the external geographic world as visualized information through gradually reducing display scale, and has become one of these core problems that occurred during establishing, processing, analyzing and expressing geographic information. As the spatial thought is confined by the multi-dimension and complication of spatial information, we should consider not only the basic spatial and attributive information of variform geographic objects but also their complicated neighboring environment which includes spatial distribution and spatial structural relation in automatic generalization of geographic information. So the generalization is not a simple combination of separate generalization of each essential but a result of multi-essential cooperative generalization. For example, there is a kind of typic restrictive relation in the generalization process between road and residential land, relief and water system. The generalization result of a kind of geographic essential usually makes some influences on that of another essential. An effective and scientific way is to construct a holistic, uniform system frame and to realize the strategy of from-above-to-below, viz. from the whole to the partial. It will help to effectively control the generalization course before producing some conflict among the geographic objects from various essentials. In order to obtain a resolve method on this problem about the multi-essential cooperative generalization, this paper will firstly think of the correlative degree of various geographic essentials, discuss their interactional relations on geographic space and attribution, then analyze the describing methods of these relations by means of mathematic, data structure and generalization knowledge, and finally research on how to realize the multi-essential cooperative generalization under the environment of computer software system.

  Being different from the traditional generalization that mainly thinks about the single essential, the cooperative generalization needs to establish those spatial and semantic relations that have existed not only in each essential but also among several essentials. Spatial relation is the most important information in the generalization, which includes the generic spatial relation, the discrete spatial relation and the spatial structure relation. The generic spatial relation is traditional topological relation that exists in the adjacent objects, and the discrete spatial relation is applied to describe the spatial contact among two or more isolated objects through TIN, QTM, voronoi graph, convex and spatial buffer method. Spatial structure relation is a very important relation on geographic meaning, such as the tree structure of water system and the network structure of road. To the three kinds of spatial relation among various geographic essentials, a key task is to resolve how to combine one relation with another during the cooperative generalization. This will lie on the types of the generalized essentials and their general relation. In this paper, a cooperative method of multi-essential generalization is put forward, discussed and practiced.