Vision-Based Measurement of Three-Dimensional Geometric Workpiece Properties

This paper documents a research program undertaken at the Industrial Research Institute of Swinburne University of Tech-nology (IRIS) in Melbourne Australia. The objective of the research was to investigate the potential for applying vision-based inspection of workpieces as a means of providing feed-back (i.e. compensation) for an industrial process. In this research, the emphasis was on high-accuracy identification and measurement of 3D properties from a 2D image. A number of techniques, for providing the transformation of the image, are documented. A common, sample process (i.e. the turning operation), and range of geometric entities were selected to demonstrate the methodology and to highlight various aspects of the proposed techniques. The paper concludes with a list of criteria that influence the accuracy of the final results.