A Hough Line Detection Based Glass Surface Defects Recognition Algorithm

This paper proposes a defect detection algorithm based on defects feature on the glass surface. The process of defects recognition can be divided into five parts. First, median filter is used to eliminate the noise jamming of the glass image. Then laplacian sharping method is applied to enhance the image contrast. After that, the image is segmented by threshold to extract defects. In order to solve the problem of the noise in binary image processing, this paper uses morphological closure operation to remove the tiny connected areas. Finally, scratches are detected by Hough transform. The results of experiment verified that, compared with traditional methods, image processing based on the proposed method has higher detection efficiency and accuracy.