An approach is developed in which sets of binary topological and directional relationships for spatial features is constructed based on definitions obtained from extending Allen's temporal relations (1) to the spatial domain. These relationships are then modeled by an abstract graph representation that preserves the topological and directional information associated with a pair of objects. Data structures are developed that, along with quantitative information stored with the graph, facilitate retrieval of fuzzy topological and directional information. A low-level query framework, including a BNF and fuzzy query processing strategies, is then presented. It is shown how this framework can be incorporated into existing high-level spatial query languages in order to provide access to the data modeling capabilities of the approach developed here
