Research On Measurement Method Of Circular Parts Based On Machine Vision

In recent years, the measurement technology based on machine vision is widelyused in parts of industrial production with the development of science and technology.The measurement technology based on machine vision which sees optical technologyand computer technology as the foundation is a modern measurement technology. Ithas a broad development prospects with the advantage of non contact, high speed,good real-time performance, moderate accuracy and strong anti-interference ability,and has caused domestic scholars of great interest in recent years. But, the existingparts size measurement method based on machine vision have high requirements tothe precision of the collection image of the parts, for the measured part image issubjected to external noise interference in the practical measuring process. In thiscase, the use of the existing parts size measurement method based on machine vision,will make the measurement system affected with the precision and stability, thusseriously affect the measurement results. Therefore, parts size measurement methodbased on machine vision has important significance.To overcome shortcomings of existing non-contact parts size measurementmethod as low measurement precision and weak anti-interference ability, a kind ofparts size measurement system based on machine vision is designed in the refer to alarge number of documents basis. The measurement system hardware part is made upof experimental stage, main USB digital camera and computer. It is suitable forwaging simulation in laboratory environment in experiment and research for thestructure is very simple; the measurement system of software design combinespowerful graphics interface function of Visual Baisic software with image processingand data calculation function of Matlab. It has used Visual mixing programming ofprinciple Baisic and Matlab, so as to realize the automation and visualization ofmeasurement process.In order to overcome the problem that outside interference image parts in theprocess of the measurement can affect the measurement accuracy and stability in circular parts measurement system based on machine vision, the basic principle ofparticle filter based on the bayesian estimation thought nonlinear filtering method isanalyzed, and a kind of optimization algorithm based on particle filter is put forwardin this paper. And a new round parts size measurement method based on machinevision, which combined with the optimization algorithm based on particle filter, is putforward in this paper. Finally, the simulation experiment in the collected digital imageby different noise interference, the method presented in this paper to get the relativeerror of measured results and variance, are far less than the existing least squaresfitting method, the pixel count method and the edge with the result of the test. It hasshows that the circular parts measurement method in this paper has high accuracy,which can meet the accuracy requirement of general industrial detection, and it alsohas good reliability and low cost advantages. This also provides a kind of way forfuture industrial machine vision to the measurement of the specific application.