A Complete Classification of Occlusion Observer’s Point of View for 3D Qualitative Spatial Reasoning

Abstract

Qualitative Spatial Reasoning (QSR) theories have applications in areas such as geographic information systems (GIS), robotics, biomedicine and spatial databases. Several region connection calculi have been proposed for use in this capacity. Primarily the existing QSR theories have been applied to 2D data. Yet the ability to perform qualitative reasoning over a collection of 3D spatial objects is desirable. Over the past two decades several theories have appeared for accurately representing and acquiring 3D spatial knowledge: LOS-14, its extension ROC-20, occlusion calculus OCC, fourteen occlusion states OCS-14, and a recent visual VRCC3D+ with 17 occlusion predicates. Each has a positive impact. However there are still some issues and ambiguities that require unambiguous ontology and resolution. In this paper, we provide a new set of self-documenting predicates for 3D complete spatial relations including occlusion. In addition we provide new heuristics for eliminating the time consuming computations by employing efficient data structures. This improvement will greatly enhance the usefulness and usability of aforementioned systems.