Quality evaluation of Automotive Body Robot Roll-Hemming Forming Process Based on Computer Vision

The quality defects in the robot roll-hemming forming process will cause the automotive body to fail to meet the requirements of appearance and assembly accuracy. To this end, a method of robot roll-hemming forming quality evaluation based on computer vision is proposed. Firstly, the quality defects of body roll-hemming forming are divided into two categories: dimensional defects and roll-hemming failure. For the dimensional defects of the automotive body, the edge detection of the automotive body and the die is carried out by the fusion of improved Canny operator and morphological open operation, and the dimensional deviation of the automotive body is measured by monocular vision; for the roll-hemming failures, the YOLOv5 model is used for real-time classification and detection. Finally, when the automotive body dimensional deviation exceeds the allowable tolerance range or the roll-hemming failure is detected, the roll-hemming forming quality is judged to be unqualified and the robot roll-hemming forming process is ended. The experiment shows that the above method can effectively evaluate the quality of the robot roll-hemming forming process.