Improved template matching algorithm based on machine vision for detecting printed code
                                         QIAN Yu hao, ZHOU Jun, TIAN Sheng, LI Shao hui
                  (School of Mechanical and Electrical Engineering, Hohai University, Changzhou 213022, China)

Abstract: Aiming at improving the detection speed of bottle cap printed code recognition in printing industry, an improved template matching algorithm based on Matlab image integral was proposed through the study of image integral algorithm and the improvement of correlation coefficient calculation formula in template matching. An experimental method was designed to acquire the image signal through the CCD industrial camera, to process the image integral processing and four check values. The pixels in the target area was determined quickly to obtain the correlation coefficients. The feasibility of the algorithm was verified by Matlab software programming test and the memory reading overhead of the original template matching algorithm was greatly reduced. The results indicate that, in terms of detection speed, the improved matching algorithm can increase speed up more than 20% compared with the traditional matching algorithm.

Key words: printed code recognition; image integral; template matching; four check values; detection speed