Research On The Key Techniques Of Glass Defects Inspection

With the transition from a glass manufacturing country to a powerful glass manufacturing country,quality inspection is particularly important in glass manufacturing.Presently,most native glass manufacturers still adopt traditional manual inspection pattern,that is one critical factor which causes the low efficiency,high cost of production,and the low quality of native glass.In this paper,the machine vision and computer image processing technology will be used to study the key techniques of the glass defects inspection system.In this paper,the main contents are as followings:The paper completes and designs the detailed design of the glass edge defects detection algorithm based on contour.Firstly,binary image is obtained through adaptive threshold segmentation,and then based on Freeman chain code tracking,the contour of glass edge is extracted,and based on the edge contour information,the inspection of glass size,edge breach,angle is completed.In order to reduce the time of contour tracking,this paper improves the chain codes tracking method.This in accordance with the direction of the search boundary point on the contour tracking method is used to reduce the tracking direction choice times and overall improves the contour tracking speed.The paper designs an ink mark recognition algorithm.The algorithm first extracts the local area of the ink marks,and then accomplishes pre-processing,connected component labeling and contour tracking.Then chooses perimeter,area,angle,circle degree and length width ratio of five geometric features to recognize the ink mask.Finally,the simple design of the glass defects inspection system was completed,and summed up the complete image processing algorithms.An application instance of the system is showed.