Robust identification of weld seam based on region of interest operation
- PDF / 2,439,261 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 42 Downloads / 153 Views
Robust identification of weld seam based on region of interest operation Ying-Zhong Tian1 • Hong-Fei Liu1 • Long Li1 • Wen-Bin Wang2 • Jie-Cai Feng1,3 Feng-Feng Xi4 • Guang-Jie Yuan1
•
Received: 17 March 2020 / Revised: 25 June 2020 / Accepted: 25 September 2020 Ó Shanghai University and Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract For welding path determination, the use of vision sensors is more effective compared with complex offline programming and teaching in small to medium volume production. However, interference factors such as scratches and stains on the surface of the workpiece may affect the extraction of weld information. In the obtained weld image, the weld seams have two distinct features related to the workpiece, which are continuous in a single process and separated from the workpiece’s gray value. In this paper, a novel method is proposed to identify the welding path based on the region of interest (ROI) operation, which is concentrated around the weld seam to reduce the interference of external noise. To complete the identification of the entire welding path, a novel algorithm is used to adaptively generate a dynamic ROI (DROI) and perform iterative operations. The identification accuracy of this algorithm is improved by setting the boundary conditions within the ROI. Moreover, the experimental results confirm that the coefficient factor used for determining the ROI size is a pivotal influencing factor for the robustness of the algorithm and for obtaining an optimal solution. With
& Guang-Jie Yuan [email protected] 1
School of Mechatronic Engineering and Automation, Shanghai University, Shanghai 200444, People’s Republic of China
2
School of Mechanical and Electrical, Shenzhen Polytechnic, Shenzhen 518055, Guangdong, People’s Republic of China
3
Shanghai Rui Rong Laser Welding Technology Co., Ltd, Shanghai 201306, People’s Republic of China
4
Department of Mechanical, Aerospace and Industrial Engineering, Ryerson University, Toronto, Canada
this algorithm, the welding path identification accuracy is within 2 pixels for three common butt weld types. Keywords Passive vision Noise reduction Welding path identification Region of interest (ROI)
1 Introduction Welding robots are very prominent in the manufacturing industry. In factories [1], most welding robots are applied to welding processes by teaching or offline programming. This method requires strict fixation and the positional accuracy of the workpiece. When the fixed position of the work-piece changes, the robot must be reprogrammed, which is time-consuming. In autonomous manufacturing, the time requirements of offline programing are reasonable for mass production, but not suitable to low or medium volume manufacturing. Therefore, to improve the efficiency of welding robots, a sensor must be incorporated. Currently, there are various different sensor types used for welding path identification, such as inductive sensors [2], arc sensors [3], ultrasonic sensors [4], magneto-optical sensors [5], and vision
Data Loading...
