Integration of Multiple Feature Detection by a Bayesian Net for 3d Object Recognition

This paper proposes a general framework to build a 3d object recognition system from a set of CAD object definitions. Various, reliable features from object corners, edges and 3d rim curves are introduced which provide sufficient information to allow identification and pose estimation of CAD designed industrial parts. As features relying on differential surface properties tend to be very vulnerable with respect to noise, we model the statistical behavior of the data by means of Bayesian nets, representing the relations between objects and observable features. This allows to identify objects by a combination of several features considering the significance of each single feature with respect to the object model base. On this basis robust and powerful 3d CAD based object recognition systems can be build.