Study On The Defect Detection Of Graphic Printing Based On Machine Vision

With the continuous development of people's living standards, presswork becomes more and more subtle with higher press speed, which meets the needs of customers with high quality and diversity. Because of the effect of subjectivity in defect detection artificially, the detect speed and precision can not be meet in industry. It is imminently necessary to take the machine vision and image processing into roboticized detection of presswork. With the development of software and hardware of computer, it is feasible to carry on the surface quality detection of presswork with machine vision. Make use of the theory, some successful manufactures have been produced in few manufacturers all over the world.Based on analyzing and studying the common machine-vision-based inspection system, an on-line defect detection experiment system for printed matter based on machine vision is designed and built, which provides a platform for studying on-line printing defect detection. This system is consisted of a computer, a line scan CCD camera, a frame grabber, a light source and a rotary encoder. With these hardwires and the data access method made up of the Halcon image processing plat and VC++, it is feasible to collect and detect colorized presswork online.In the system of software, presswork manuscript template establishing has been achieved by gray-value-based template establishing. The software matches the detected image and the manuscript template using image pyramids, and it compares the detected image with the manuscript template with difference image method. The system can carry the point of match and detect fast.