A multi-sensor approach to automating co-ordinate measuring machinebased reverse engineering

Co-ordinate measuring machines (CMMs) have been widely used in industry for reverse engineering. In an ideal reverse engineering system, the object would be automatically digitised, the data would be segmented into constituent surface patches, and an accurate solid model would be generated. However, at present, the time-consuming manual digitisation of many objects discourages the implementation of reverse engineering. As a major step toward attaining this goal, a charged coupled device (CCD) camera and a CMM touch probe digitiser are used together in a novel reverse engineering process. Neural network based stereo image processing is used to locate the object in the CMM work space, and to generate the CMM touch probe path. Precise data is then collected using the CMM touch probe, from which a CAD/CAM model can be constructed.