Seam detection method design and automatic calibration based on an active light coaxial visual sensor
Zhehao Zhang,Yi Zhang,Cong Chen,Xiangyu Wang,Jie Li,Wei Xie
DOI: https://doi.org/10.1016/j.optlaseng.2022.107304
IF: 5.666
2023-01-01
Optics and Lasers in Engineering
Abstract:Seam feature detection is one of the critical premises of automatic welding. However, the current seam detection method is vulnerable to complex laser environments, such as illumination interference and material defects, thereby limiting the application of active light coaxial visual sensors in seam tracking. This paper proposes a seam detection and calibration method based on an active light coaxial visual sensor in a laser welding digital twin environment. First, a vision sensor with an 808 nm auxiliary illumination and a customized lens are mounted on the laser welding head to filter out noises and capture the welding image, including the seam and melt pool. Next, a seam information extraction approach based on image processing algorithms is designed to extract seam position and seam width. In addition, an artificial neural network based Kalman filter is proposed to eliminate measurement errors and reduce measurement noise. The calibration of the seam detection parameters is a tedious process. Therefore, the calibration problem is transformed into an optimization problem. A genetic algorithm-based calibration method is proposed to automatically determine the optimum parameters of the seam detection algorithm. Finally, seam detection experiments on laser welding are conducted in both the digital twin environment and the laser welding system. The experimental results show that the proposed method is more efficient than the Hough transfer-based seam detection method. The calibration method can ensure the accuracy of the seam detection method, especially in complex and changeable welding conditions.
optics