Integrating Multiple Data Representations For Spatial Databases

An intelligent spatial data,base must be able to organize and store information from diverse sources. Aerial imagery, map, and terrain data must be merged with textual and collateral information. Future systems will integrate the results of automated analysis of remotely sensed imagery within the context of the spatial database. No single internal representation can efficiently provide for the variety of the needs and problems for these types of spatial databases. For example, in order to efficiently search large databases it is critical to be able to partition the search based on spatial decompositions, whether hierarchical, regular, or mixed. In this paper we discuss some recent work on integrating multiple spatial and factual data representations so as to capitalize on their inherent advantages for search, geometric computation, and maintenence of topological consistency. 1