Automatic Vision Inspection System for Printing Registration Accuracy

Registration accuracy is one of the important performances indictor for the quality of printing. In order to overcome the disadvantages of traditional manual to inspect register accuracy lead to slow efficiency and low precision, an automatic vision inspection system was designed. Register crossline mark images were collected. The crosslines were segmented by FCM algorithm. Based on Hough transform, a method was proposed to extract the edge features of crosslines. The system was visually calibrated and the distances between crosslines were measurement to get the registration accuracy. MATLAB GUI function was used to design system software. Results show that this system having high detection accuracy is feasible.