Spatial Topological Queries

In geographical information system(GIS), as an important part of spatial relation, topological relation is the base for constructing of GIS database, storing and querying of spatial character, and topological relation provides the base for spatial analysis too. Hence, the study of the spatial relationship has important academic and practical scene in advancing the development of GIS. With popularization of the GIS application, spatial database has been widely used in various departments. As a basic function of spatial database application system, spatial topological query receives increasing attention and becomes the standard of spatial data management capabilities in spatial data management system. This paper mostly studies and discusses refining upon 9-intersaction model and designing topological query application.9-intersaction model is the most mature developed and most widespread topological relation model, it describes topological relations qualitatively, therefore the description is a rude classification of relations between entities in real world. This paper introduces the metric parameters, refines the categories of topological relations identified by 9-intersaction model via using the method with the combinations of topology and metric to describe topological relations, and designs topological query application with support to metric parameters.Spatial topological query is divided into two categories :(1)Find the objects that satisfy particular topological relation with given object, namely determine whether two objects satisfy particular topological relation. (2)Queries on the topological relation between two objects. Existing decision tree building algorithm assumes that all topological relations occur with equal probability, but this situation is extremely rare in practice. This paper refines the existing algorithm via assuming occurrence frequency of topological relation. Comparing with the decision tree built by the existing algorithm, the decision tree built by the refined algorithm has a lower weighted path length(WPL), and more closer to optimal decision tree.