Detection of characters ink dot defects on the surface of industrial products
Xingsi Liu, Min Xia, Nanqing Xia, Shizhen Tang, Aohan Yao, Wei Li · 2021 4th International Conference on Advanced Electronic Materials, Computers and Software Engineering (AEMCSE) · 2021
In order to solve the problem of long time-consuming and poor adaptive ability of the current character printing quality inspection system, a character quality detection system based on machine vision is proposed for the inkjet printing method. Prior knowledge is used to divide the entire character area into individual character areas. Then template matching is carried out for each single character area. The template matching result is further operated to find the inner area of the character.The inner area of the character refers to the area that should be printed by inkjet printing. The dot area detection method is used for the inner area of the character to detect whether there are missing ink dots or blurred ink dots, as well as inaccurate ink dots. The detection system has good performance under uneven illumination and machine jitter conditions in actual production line. The detection speed of a single character can reach within 15ms, and the detection speed of a single sample can reach within 300ms. After experimentation, the comprehensive detection success rate of the three defects is 86.68%, and the false detection rate reaches 0.71%.