Welding component identification and solder joint inspection of automobile door panel based on machine vision
Zezhong Pei, Liding Chen · 2018
By applying machine vision technology to automobile door welding production lines, a welding component identification method based on Canny Edge Detection Algorithm and SSDA (Sequential Similarity Detection Algorithm) is proposed in this paper, and the image segmentation is used to improve the accuracy of Hough Transform in inspecting solder joints. Firstly, the Canny algorithm is employed to get the edge image of the automobile door panel, Then, the SSDA is applied to match that edge image with template images of welding components to obtain the number and types of the welding components on the automobile door panel. Finally, the edge image is divided into several target image regions according to the identified welding components, and the positions and number of solder joints in each segmented image region are inspected by Hough Transform respectively. Experimental results demonstrate the effectiveness of this method.