Spatial queries and data models

Abstract

We present a unified framework for classifying and answering spatial queries relevant to a Geographic Information System. We classify spatial queries into topological, set-theoretic, and metric queries, on the basis of the kind of relationships between the query object and entities in the search space involved. For answering such queries, we propose an approach that combines an object-based description of spatial entities, provided by a topological model, with a partition of the space embedding such entities, given by a spatial index. In particular, we propose a new unified topological model, called the Plane Euclidean Graph (PEG), that is capable of describing point, line, and region data, and that incorporates relational operators on such entities. We briefly describe major techniques, rooted in computational geometry, for solving interference queries and overlays on such a data model. Finally, we describe the use of a superimposed spatial index for speeding up searches and answering queries involving distances.