Study Of Spatial Analysis Method Based On Multi Spatial Granularity

Spatial on-line analytical processing is an important foundation for spatial data mining in spatial data warehouse. However, as the main method of spatial data mining, spatial analysis gets very limited support from spatial on-line analytical processing. To solve this problem, the division of spatial granularity in spatial data warehouse is discussed in this thesis, and on this basis a spatial data warehouse based on multi spatial granularity is built combined with the technologies such as spatial data cube, spatial measure pre-computation, etc.First, this thesis introduces the definition, characteristics and architectures of spatial data warehouse briefly, together, the basic concept and structure of spatial data cube is described. Meanwhile, the spatial data cube operations and the spatial measure pre-computation is presented.Secondly, this thesis discusses the division of spatial granularity in spatial data warehouse as the basis for the design of spatial data warehouse model based on multi spatial granularity. Moreover, spatial analysis methods based on multi spatial granularity is proposed, which focus on the analysis of spatial topological relations and the overlay analysis. After that, as the most frequently used operation in the spatial on-line analytical processing, merging a set of polygons is studied, and an improved algorithm named grid cascade union is put forward by analyzing the inadequacies of the existing algorithms.Finally, in order to verify this study, a spatial data warehouse prototype based on multi spatial granularity is built in this thesis with a topic of land use. Performance testing and application shows that, this study can effectively improve the speed of spatial analysis in spatial on-line analytical processing as well as the efficiency of merging a set of polygons, with strong practicability.The results of this study have some theoretical and practical guidance to the construction of spatial data warehouse and the application of spatial on-line analytical processing.