Glass Defect Identification Method Based On Image Processing Technology

With the development of machine industry, detection of glass defect has being faced with more opportunity and challenge. Traditionally, the judging work is done by judging worker, the man analyses many defects of glass and identifies the sort of defect. The traditional method has many disadvantages, for example, bad effect from man's quality, low efficiency and so on. It does not meet requirement of productivity gradually. So new judging technologies of glass defect have been presented. It is intelligent judging technology through computer.Solved the digitization of glass image, digital image processing become an important part in the identifying glass defect. Many researches about the glass image have been done in this paper. These researches are preprocess of image, analysis of image, extracting defect's characteristic information and classified defect. Preprocess of image is the base of following image processing. After analyzing image's noise model, effective filtering method is presented in this thesis. Aims at the glass image's character of low gray contrast and faint edge, linear transformation is used in the article. It stretches the gray scale, enhances the gray contrast and clears the image. Half threshold value division is used to erase the shadow of defect in image because of experimental condition limitation.Selection and extraction of defect's characteristic parameter are precondition for defect classification, it directly affect the result of classification. Studied character of defect, a system of glass flaw characteristic parameter has been presented. The identification of glass defect is multiple objective identification. In this kind of situation, it is essential to mark every defect. The article proposes a method of numbering mark which solves the deficiency of gray mark'range perfectly. Defect mark is the effective method to select defect to operate. The article has analysis flaw selective outline track based on marking.The above-mentioned solution to intelligent defect recognition of glass has passed test and has a perfect effect.