Research Of Steel Ball Surface Defect Detection Technology Based On Machine Vision

China is a country of many ceramic tile production and production enterprises.The ceramic tile surface quality not only reflects the beauty of the surface,but also affects the performance of the ceramic tile to a certain extent.Therefore,ceramic tile surface quality detection has become an indispensable part of ceramic tile production enterprises.With the rapid development of machine vision technology in the field of industrial detection,machine vision detection technology has become an important research direction of ceramic tile surface quality detection.On the basis of the deep understanding of the field of machine vision and the related algorithms of image processing,this paper carries out the research on the surface quality detection technology of ceramic tile,taking the surface defect of ceramic tile as the research object,the algorithm of ceramic tile surface defect detection is taken as the core to achieve efficient detection of ceramic tiles with different defects,so as to achieve accurate classification.Through the research on the edge detection operator analysis,a local variance rotation invariant measure edge detection algorithm combining with close operation of mathematical morphology is proposed in this paper.The experimental results show that the proposed algorithm can effectively obtain the tile surface defect image edge detection and get high accuracy rate.The main contents of this paper are as follows:(1)The design of the ceramic tile surface quality detection system.The hardware part includes selection of camera and lens,data output interface,the selection of the light source and the design of the lighting mode,etc.The software part includes VS2013 software and Halcon10.0 configuration problem,detection algorithm,defect marking,feature extraction and defect image classification algorithm,etc.(2)The preprocessing of ceramic tile surface defects image.Aiming at the ceramic tile surface image obtained by CCD linear array camera,the possible noise is researched,and get the complete information of the image through median filtering,de-noising,histogram equalization,image enhancement and other preprocessing processes,so as to lay the foundation for subsequent defect detection and classification.(3)Based on the image segmentation of ceramic tile surface defects,a local variance rotation invariant measure edge detection algorithm combining with close operation of mathematical morphology is proposed.Firstly,local variance rotation invariant measure operator is applied to detect the surface defect image of ceramic tile,and the image is converted to binary image by setting threshold.Secondly,the new image matrix is subtracted from the detected image and the defect free template image matrix,and the median filtering is adopted again to remove the introduced noise.Then the ceramic tile defect image is detected by the closed arithmetic operator of mathematical morphology,so as to obtain the effective edge of the defective image.The experimental results show that the proposed edge detection algorithm completely preserves the edge details of the defect area,and achieves the accurate location of the defect edges,and the detection efficiency and accuracy are relatively high.(4)According to the ceramic tile surface defect image detection,further identification and analysis such as defect marking and the feature extraction process should be done,using support vector machine(SVM)to classify the different defect types of ceramic tile,using cross validation method to realize the high-precision classification of ceramic tile surface defect images.Through the analysis of detection and classification experiments of ceramic tile surface defect image,the optimal parameters are determined in the experiment process.And other algorithms were compared to verify the superiority of the algorithm from different angles.