Hierarchical Bayesian methods for recognition and extraction of 3-D shape features from CAD solid models

This paper introduces a new uncertainty reasoning-based method for identification and extraction of manufacturing features from solid model description of objects. A major difficulty faced by previously proposed methods for feature extraction has been the interaction between features. In interacting situations, the representation for various primitive features is nonunique making their recognition very difficult. We develop an approach based on generating, propagating, and combining geometric and topological evidences in a hierarchical belief network for identifying and extracting features. The methodology combines and propagates evidences to determine a set of correct virtual links to be augmented to the cavity graph representing a depression of the object so that the resulting supergraph can be partitioned to obtain the features of the object. The hierarchical belief network is constructed based on the hypotheses for the potential virtual links. The evidences which are topological and geometric relationships at different abstraction levels impact the belief network through their (amount of) support for different hypotheses. The propagation of the impact of different evidences updates the beliefs in the network in accordance with the Bayesian probability rules.