Combining ontological and qualitative spatial reasoning: Application to urban images interpretation

Abstract

In this paper, we develop a qualitative spatial reasoning (QSR) applied to the interpretation of urban satellite images. This reasoning is integrated within an ontology of urban objects, which has been implemented in cooperation with expert geographers. The spatial concepts mainly include the set of topological relations of the Region Connection Calculus theory (RCC8), and the set of computational primitives (CM8) which have been defined for computing the RCC8 relations on raster images. Our approach relies on a reified representation of the RCC8 relationship and of the CM8 primitives, within a lattice of concepts, implemented in OWL (Ontology Web Language). It provides a straightforward representation of concepts corresponding to conjunctions or disjunctions of spatial relations, and thus offers the advantage to overcome some drawbacks of the existing approaches in OWL, where spatial relations are represented as roles. Indeed, the OWL language does not allow the expression of the disjunction of roles. We can then implement a reasoning on the RCC8 relationship, which in particular allows the computation of the composition table and its transitive closure. As the reification of roles precludes the use of role’s properties, such as symmetry and transitivity, we propose to implement RCC8 inferences through SWRL rules (Semantic Web Rule Language).