Research on the application effect of machine vision technology based on simulation analysis in 3D inspection of torque converter welds

Abstract

With the development of machine vision technology, visual inspection technology has been widely used in many fields. Among them, the 3D visual detection technology based on line structured light has attracted the attention of relevant scholars because of its advantages of the easy acquisition of image information and high detection accuracy. Therefore, applying machine vision technology to the field of three-dimensional weld inspection can improve the inspection effect, so a machine vision technology based on simulation analysis was proposed. First, the model parameter calibration and algorithm of the structured light imaging system are studied, and then a structured light-based torque converter weld height detection system is designed and related models are constructed. Finally, the detection accuracy and speed of the constructed 3D vision measurement system are verified. The results show that the calibration parameters obtained by the optimization algorithm of this study are as $t_1$ follows, focal length = 17.16376; $u_0$ = 1798.55439; $v_0$ = 1276.85505; $r_1$ = 0.14072; $r_2$ Secondly, in the comparative experiment, the height of the wave crest coordinates of the measurement results of the three-dimensional detection system of the torque converter weld seam in the global feature extraction process based on the machine vision technology based on simulation analysis is 9.1330 mm, and the maximum error value is 1.0403 mm; As a result of the calibration algorithm, the peak coordinate height is 10.3526 mm, and the maximum error value is 2.7482 mm. The error of the research system is better than Zhang Zhengyou's calibration method. The research content has important reference value for improving the three-dimensional inspection effect of torque converter welds.