Automated reasoning on vague concepts using formal ontologies, with an application to event detection on video data

In this paper we lay the foundations of a formal ontology for the characterisation of vague concepts sourced from natural language, applying principles of event calculus and super-valuation semantics. We focus on a specific set of motion events and related concepts, motivated by the aim to develop an automated reasoning system able to detect occurrences of such events in video scenes. Our goal is to provide a general methodology for the formalisation of vague concepts, and to address the issue of vagueness in formal ontologies.