Research on hole feature matching based on image restoration
Yi Wang, Zhengdong Ma, Hao Su · 2022 International Conference on Computers, Information Processing and Advanced Education (CIPAE) · 2022
Aiming at the problem that stains and strong light reduce the accuracy of light hole and threaded hole matching in the process of traditional template matching, a template matching method for body in white hole feature based on image restoration is proposed. Firstly, the patch match repair method is used to repair the image of stains and strong light, so as to eliminate the interference of the two to the matching. Then the light hole and thread are matched by combining the gray matching method of improved search strategy and the edge feature matching method respectively. Finally, the sub-pixel coordinate fitting is performed on the matching results. Experimental results show that the accuracy of matching smooth holes and threaded holes can reach 100%and 98%, and the difference between matching accuracy and commercial software matching accuracy is less than 0.3.