Formal specification of region-based model for semantic extraction in road traffic monitoring

Abstract

This work forms part of the development of a framework for semantic extraction in road traffic monitoring. In this paper we develop a scene, object and event model based on regions in the ground plane. The model is formally specified using the Güting spatio-temporal formalism for moving regions and Z notation. The result is domain-independent knowledge representation that supports reasoning about time-varying regions and that is expressed in an accessible mathematical formalism.