EFFICIENT DATA REPRESENTATION AND SPATIAL RELATIONS COMPUTATION

Abstract

Data representation and spatial relations computation are the essential basis of spatial information systems (SIS). In this thesis we present a new data structure, the dual data structure, to efficiently represent the spatial data in SIS which integrates both vector and raster format data. We utilize a new concept, visual boundary, to describe the boundary between region objects. On the basis of the visual boundary and dual data structure we give a consistent set of definitions for various kinds of spatial relations, including direction relations, topological relations and metric relation. This set of definitions can be used as a solid basis for spatial reasoning and for developing a spatial query language.