The Design Of Hole Parts Detection System Based On Computer Vision

Detection is one of the basic activities of industrial production and manufacturingprocesses. Quality information can be obtained through the detection of product and itsmanufacturing process. According to the information on the manufacturing process, wecan take correction and compensation to make the rate of scrap and rework reduced tothe lowest level to ensure the quality of output products. With the progress of society,the continuous development of science and technology, the measurement and researchof micro parts industry increasingly high demands. But the traditional detection mainlyrely on people, and mainly rely on manual completion. It is both time consuming andcost, the time of production growths, the production cost increases, and not theimplementation of real-time online detection or control in machine process.In view of some of the drawbacks of the traditional method of tiny parts detection,detection method based on computer vision is proposed in this paper. Detectiontechnology based on computer vision, subjective judgment does not depend on thepeople, but by a computer and corresponding software technology to complete theinspection of products. With the help of appropriate software, it can complete the task ofdata acquisition, processing, feature extraction and recognition automatically, and avariety of analysis and calculation. It is flexible, fast, intelligent, real-time, in manyadvantages of linear etc.This paper focuses on the detection problems of circular holes in micro mechanicalparts. The mechanical parts of the every kind of shape, and the application scope is veryextensive, especially the realistic significance of the study on slight holes in mechanicalparts, and it also has certain guiding significance to production and installation ofmechanical parts. This paper selects the transmission chain plate with two holes, theoriginal image acquisition, image smoothing, image segmentation, edge detection,combined with the specific algorithm to extract the center, calculate the radius ofcircular hole and the distance of the two holes.Through theoretical analysis and experiments, it is show that the proposeddetection system based on computer vision can successfully detect small parts of the shape and size, and has higher measuring accuracy. It has the important practicalsignificance in large-scale industrial production.