How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Advanced Machine Vision Surface Inspection
Thanks to modern quality control techniques such asTaguchi's Loss Function, many manufacturers now appreciate the cost savings and quality improvements attainable via reductions in product variation, even when variations are within a specified tolerance...
By Lyndon N. Smith and Melvyn L. Smith
One bottleneck in the effective implementation of SPC is the availability of real-time accurate measurement data from the production line. This is particularly true for surface quality control, where large areas have to be measured or inspected and the available technology has proved to be quite expensive or unable to fully address the needs of the task.
This has lead manufactures in many sectors to rely on manual inspection, resulting in subjective and variable results. Machine vision is providing a means for changing this situation through the application of high-res and low-cost imaging and analysis techniques, thanks to research currently underway in the Machine Vision Laboratory at the University of the West of England.
There are strong cost and productivity motivators for realizing effective automated surface inspection. For example, in 1998 it was reported that within a US automotive paint inspection research program, the time taken to recover the R&D expenditure was under one month and the payback period for a complete retrofit of inspection was under six months [Lloyd 1998]. The importance of being able to image specular surfaces such as car bodywork has lead to a considerable amount of work on the analysis of specular and diffuse reflections from painted metal surfaces [Parker, et al. 2002].
At the same time, industrial applications in this area continue to be developed, with an example being provided by a major steel manufacturer based in the UK [Manufacturingtalk.com 2002]. Regarding vision analysis of bare metal surfaces, research reported now regularly describes the use of artificial intelligence techniques for defect detection. Examples include an Australian university that is researching into genetic algorithms [Zheng et al. 2002] and work in Finland on neural networks for real-time inspection of steel strips [Machine Vision News 2002].