Inferring complex geographical concepts with implicit geometries using ontologies: A case of peninsulas

Abstract

Ontology-driven concept inference has the merit of high flexibility and transparency. Users can composite and reuse atomic primitive concepts and relationship to interpret complex geographical concept without the need to retouch or even know the technical details. The major issue that we are focusing on is the implicit geometry problem. That is, the geometries corresponding to some primitive concept defining the complex geographic concept are missing or not fully represented in a spatial database, making it impossible to inferring the high-level semantics of the objects. This paper combines terminological/assertional inference (for general logic reasoning) and spatial operations (for making implicit geometries explicit), therefore enabling an ontology-driven inference of complex concepts that can handle cases where some concept has no explicit geometries. In the end, the concept of peninsula is used to demonstrate the proposed methodology.