An Efficient Representation of Qualitative Spatial Information using Bintrees

Abstract

We present a representation of spatial data that supports answering of qualitative queries (involving e.g. containment and overlap), using less time than numerical computations on coordinates, and less space than explicitly storing the relations. The representation is based on the well-known bintree indexing data structure, which we refine to provide exact answers instead of approximations. We have imple- mented the method, and tested it on several real-world datasets with promising results.