Bearing Defect Detection Based On Machine Vision

The quality issues of bearing is always highly concerned,because bearing has been widely used in daily life as a component,however,defects such as bumps,scratches,and rust are appeared on the bearing surface during the production process,and the phenomena of bearing rollers missing is also present.If manual defects are used to detect bearing defects,not only the speed is slow and the efficiency is low,but also the detection accuracy is affected by many factors.Machine vision technology has been widely used in many fields,based on its fast,flexible,and strong anti-interference capabilities,which can achieve the fast and accurate positioning and recognition of different detection tasks.For this,the algorithm of machine vision is introduced in this paper to achieve the detection of bearing surface defects and roller missing assembly defects,to improve the speed and accuracy of bearing defect detection,and to classify bearing defects.The specific research content is:First of all,a new type of Canny bearing surface defect detection method was studied.Pre-process the source image through the gamma correction algorithm,and the high and low thresholds of the traditional Canny algorithm are adaptively selected based on the iterative threshold segmentation method and the Ostu algorithm to improve the integrity and accuracy of the segmentation of bearing surface defects;on this basis,the circularity,area,perimeter,and aspect ratio of the image are used as the characteristic parameters,the support vector machine multi-classifier is used to achieve accurate classification of bearing surface defects.Secondly,a detection method based on edge template matching and Hough bearing roller missing assembly defects was discussed.The homomorphic filtering algorithm is used to enhance the contrast of the image,and the Hough transform algorithm based on edge template matching is used to obtain the contour of the bearing roller,to reduce the impact of the image illuminance,and find the original image area where the center position of the bearing roller is located;then the edge template matching algorithm is used to screens these areas to obtain the precise position of the assembled roller,and the least squares fitting algorithm is used to perform circular fitting on the center point of the assembled roller to achieve accurate positioning of the missing roller of the bearing.Finally,a bearing defect detection system and its architecture based on machine vision are designed.The hardware configuration of the bearing inspection system was discussed in detail,and the upper computer interface of the bearing defect inspection system was developed,and the algorithm program of the bearing defect inspection was written,and the effectiveness of the system was verified through experiments.In summary,theoretical and practical application value of the study for the application and promotion of machine vision technology in bearing defect detection is important.