Vision-Based Gauging System for Dimensional Measurements

Dimensional measurement is often required in machinery manufactory industry. Vision-based gauging, as an emerging non-contact metrological instrumentation, can greatly facilitate reliable evaluations for their machinery products. In this paper, such a VGS (Vision-based Gauging System) was implemented for dimensional measurements and its industrial feasibility was verified. Both the systematic performance and potential risk factors were discussed carefully, from which it can be seen that lens distortion, imperfection in state-of-art image sensors, implicit effects of edge detection and curve fitting unexceptionally contributed to the systematic error, and the largest proportion of the systematic error herein can be attributed to lens distortion. Based on experimental results, it showed that the measurement uncertainty of VGS was about 10 micron. Having the error sources targeted, we would expect to enhance the systematic performance in future.