Research And Design Of Bearing Seal Cover Defect Detection Based On Machine Vision

As an important part of bearing production line,the defect detection of bearing sealing cover directly affects the safety and delivery rate of bearing.In order to avoid the shortcomings of low efficiency,high cost and low detection accuracy,this paper combines machine vision technology with bearing seal cover defect detection,and designs a set of bearing seal cover defect detection system based on machine vision.The system can not only improve the detection precision,shorten the detection time,but also reduce the labor cost.The main contents of this paper are as follows:(1)In the process of taking the image of bearing seal cover,the noise and fuzzy problems exist in the image obtained due to the interference of surrounding environment.Through image graying to reduce the redundant information contained in the image and reduce the amount of calculation;using image enhancement method to improve the clarity of the image,and then through the image filtering technology to remove the impact of noise;after using image binarization and morphology to highlight the characteristics of the target area,facilitate the subsequent feature extraction and defect detection.(2)The center and radius of the bearing are obtained by Hough transform circle detection algorithm.Based on the center and radius of the bearing,the ROI region of each part of the seal cover is extracted by using the correlation algorithm,and then the pixel line feature map obtained is de disturbed.According to the analysis of the defect characteristics of the seal cover,the character area defect in the middle area of the seal cover is detected by using the character recognition technology,and the non character area defect is detected by the traditional manual setting threshold.(3)Aiming at the problem that the inner and outer lip of the sealing cover is easy to be detected by traditional methods,this paper adopts SVM algorithm in machine learning to detect defects.Through the comprehensive use of low variance feature removal and feature selection algorithm based on tree model to extract features from the original feature data set,then the classification decision function of SVM algorithm is improved,and a suitable classification threshold is found by using precision and recall rate.The experimental results show that the improved SVM algorithm can detect the defects in the inner and outer lip of the sealing cover.In this paper,the 6003 bearing seal cover produced by the cooperative enterprise is taken as the test object,and the detection algorithm is used to test.The test results show that the detection accuracy can reach more than 98%,which can meet the requirements of field detection.In addition,the detection method in this paper is applied to the design of the system software,including image acquisition module,image processing module,feature extraction module and defect detection module,which achieves good human-computer interaction function.