Machine vision for 3D mechanical part recognition in intelligent manufacturing environments

Abstract

A 3D machine vision method used in intelligent manufacturing environments is presented. In this method, the neural network technology is used to provide effective methodologies for solving difficult computational problems in 3D recognition processes. The recognition processes can be divided into two parts. First, a 3D reconstruction. approach based on wavelet analysis is presented. The stereo matching problem is solved with a wavelet analysis. The dyadic discrete wavelet analysis is adopted in this process and a stereo matching process is realized with global optimization. A coherent hierarchical matching strategy is constructed, so that the stereo matching process can be accomplished with coarse to fine techniques. A 3D reconstruction neural network is constructed by using the BP neural network. With the results of stereo matching, the 3D shape of part can be reconstructed. Then the feature vectors of 3D parts are constructed by using the 3D moment and its invariant. An ART2 neural network is adopted for the neural network classifier, by which the 3D parts can be recognized and classified. The method was tested with both synthetic and real mechanical parts in intelligent assembly system. Results show that the method presented is effective and suitable for an intelligent assembly system.