The Research And Application Of Detection Algorithm Of Hole And Sides Of Defects Based On Machine Vision

Appearance quality of electronics products has become an important factor in consumer choice of electronic products, with the surface quality of the products have become increasingly demanding. However, traditional manual visual inspection method is not only inefficient, but also labor-intensive, how to quickly, effectively and intelligent to detect surface defects such components have far-reaching significance for enterprises and maintaining corporate reputation. However, existing foreign-related detection equipment is expensive and not suitable for detecting defects and side sound hole periphery defectsHole and the sides of cover glass defect detection is studied on this paper base on machine vision. The overall detection system was designed in this paper for the defect of hole and sides, including an image acquisition sub-system and the image processing sub system.According to the surrounding collected defect image feature and defect position, the defect is divided into two types of holes around the perimeter and around. Described its causes by typical defect map of the defect and analyzed the characteristics of the defect, and descriped the difficulties of solving such defects, clearing direction of later defect detection algorithms.Image segmentation method is proposed based on the characteristics of the soundhole contour to solve the classic dark vents segmentation algorithm for image segmentation problems, based on the mean gray of area of interest for the detection of defects highlights. For Spurs, rat-bite dark defects, detection method based on the soundhole contour distance to detect defects in local minima was proposed. For peripheral defects, detect the highlights deficiencies by excluding normal highlights through geometric positional relationships between highlights. For dark gap defect, according to its gray values with local characteristics of the low-frequency side of the gray values in the normal distribution curve, based on the local minimum value to detect there is no dark defects.The proposed algorithm on this paper was successfully applied to the surface Defect Detection System of the cover glass, and the test results verify the practicality of the detection algorithm. Currently, prototype has been manufactured, being in the preparatory stage of industrialization, have practical value.