Color printed image defect detection based on the image feature match

摘要

Color printed image defect detection algorithm based on the sift method and YIQ model is proposed. At first, pre-treatment is conducted with median filter for the color printing image. Feature points of color printing image are extracted with Beaudet algorithm, and then, they are described with sift algorithm. For each sift feature point, YIQ color feature is extracted. Finally, the feature vector about sift and YIQ feature of the feature points in the color printing image and standard specimen page image are matched, for determining the defect region of the color printing image. According to the experimental results, this method can detect the defects of color printing image rapidly and effectively.