A representation for databases of 3D objects

The real world is inherently three dimensional. Therefore it is natural to represent it in 3D. Many application domains such as architecture, urban planning, epidemiology, and virtual reality require representation of objects in three dimensions. These applications require an effective way of querying three dimensional data. Querying 3D objects based on their exact geometry has proven to be computationally expensive. This paper presents a representation that approximates the geometry of 3D objects. This representation is based on 3D line segments. We compare the proposed representation with the polyhedral representation to detect collision between two objects. Our experiments show that our representation can reduce significantly the computation time to detect collisions. We also show how our representation combined with meta-data can be used to effectively query 3D objects.