Constraint-Based Reasoning in Geographic Databases: the Case of Symbolic Arrays

Symbolic arrays are hierarchical constraint-based representations that preserve direction relations (e.g., north, northeast) among the distinct components of complex spatial entities. They have been used in problems involving pattern matching and spatial information retrieval. In this paper we demonstrate how inference can be achieved in geographic databases of symbolic arrays using composition of direction relations. In particular, we distinguish two types of spatial inference: the first is concerned with the inference of constraints between objects that exist at different levels in the hierarchy, while the second type involves the inference of constraints between objects that exist at the same level but in different arrays.