| With the development of president science and information, Printing products are an important part of human life which take a more important position in daily life. Accordingly the quality of the printing matter is increasingly concerned. The traditional manual method of image quality inspection can not satisfy the requirement of the inspecting speed and precision standard because of subjective factor of human being. With these reasons, introduction of image processing technology into automatic printing inspection becomes an urgent problem. With the development of computer software, hardware and the perfection of computer vision theory, the method for inspecting the image quality with computer is becoming available. In foreign countries some factories have made out this kind of product and its quality is good. Our country is researching this kind of product that has not make out any kind of finished product.Aimed at the automatic inspection of industrial printing image, this thesis discusses and studies principle and algorithm of defect inspection based on machine vision, and then realizes the function of defect inspection ability of printing quality inspection system.In the aspect of image pre-processing, methods of image quality detection have been analyzed and concluded, a scheme of realizing image pre-processing using image technology is built.In the aspect of defect inspection, according to the defect character of industrial printing image, a defect inspecting method of sampling and layering based on resolution technology is introduced. In this method the person's vision character and industry inspection request are considered, a dynamic threshold and dot-surface defect inspecting algorithm with layers are given. At the meantime, inspecting defect with layers using inspection windows with different size, the pyramid algorithm has been improved and simplified. It's an effective method to improve the precision and speed of defect inspection for printing image. |
PDF Full Text Request