Constraint relaxation on topological operators which produce a null value as answer to a query

In the field of spatial information systems, and especially in particular of geographical databases, many authors have studied how to formulate queries using pictorial configurations. The constraints deriving from topological relationships between pairs of symbolic graphical objects (the classic shapes point, polyline, and polygon) can be relaxed when a query search condition finds no match in the database, so that users can receive approximate answers rather than null information. In this paper a computational model for the similarity of the patial relations is proposed. An operator conceptual similarity (OCS) graph describes the model, and links the more similar relations, defining the weight of each relaxation. The polygon-polygon, polyline-polyline, and polygon-polyline cases are discussed and matrices, which enlarge the known 9-Intersection model matrix are also considered.