Research On Multi-surface Defects Inspection System Of Magnetic Ring

Magnetic rings are widely used on audio equipment, some surface defects will directly influence its effect, so the detection of its surface defects became the important process during the production. Machine vision detection technology is one of the research hotspot in recent years, it has the advantages of more faster, non-contact and higher precision. In this paper, Image processing technology is used to develop an on-line multi-surface defects detection system of magnetic ring based on machine vision.At first, the background and the significance of the magnetic ring defect detection system are discussed and the necessity of the application of machine vision inspection system. Then the concept and common algorithms of digital image are taken a brief overview. Followed, the vision system’s working principle and hardware selection is analyzed, including the control system of magnetic ring. Next, the control system is designed based on PLC in order to realize the control of the step motor, vision system and the communication with PC. Then, the image processing algorithm is designed according to the characteristics of magnetic rings’defects. In the image preprocessing, using the method of replacing the pixel gray value by near average value to improve the median filter, making up for the fuzzy details by median filter. Then, in order to get high accuracy of measurement, sixteen chessboard images are used to calibrate the camera, correct the distortion and get the transformation matrix. The canny operator is used to get the edge information of the ring, followed using the least squares to fit the circle parameters and get the diameter data. Before detection, the defects are divided into two kinds, first one is the defect of larger area, this paper uses the circle parameters and ROI setting to improve the subtraction algorithm and it’s efficiency. On the other hand, the small defects’ detection like cracks are more complex, first, morphology method is used to remove the influence of veins, second, using the distance calculation to remove the boundary point on binary image, third, using the feature extraction algorithm to scratch the defects image from the background, at last, a calculation method of minimum bounding rectangle is designed to judge the crack size based on a new field search method, also snake model is discussed on defects marking.The multi-surface defects inspection system which this paper designed has the advantages of high efficiency, reliable and high precision etc, also it solves the problems of low efficiency and accuracy caused by artificial detection and has high social efficiency.