An informed strategy for matching models to images of fabricated objects

摘要

A new method for the matching of 2-dimensional scenes of single and multiple objects with a previously formed model database is presented and discussed. The system is based on a two-tier organisation of the model database consisting of object centred and binary relation centred structures. This is formed to reduce the complexity of the matching process yet retain a robust and accurate identification procedure, in the event of an incomplete and inaccurate segmented description of the scene, which is a probable consequence of non-idealised viewing conditions. Examples of system performance in identifying single and multiple component images are presented.