Development of Mechanisms for Automatic Correction of Industrial Complex Tools in the Preprocessing of Laser Welding for Small-Scale and Piece Production Using Computer Vision

Abstract

This paper is devoted to the development and improvement of mechanisms for the functioning of an automated system for correcting the points of the planned trajectory of the tool of a laser robotic welding complex in the pre-process. Correction of the points of the planned trajectory is carried out in two stages: in the first stage, the focal laser radiation is corrected; in the second stage, the position of the tool is corrected. Correction of the focal laser radiation is carried out in conjunction with the automated focusing of the camera by moving the tool of the industrial complex along its own axis. The functioning of position correction mechanisms is based on methods for recognizing the edges of the gap line to be welded from the image obtained from the charge-coupled device (CCD) camera. The edges of the gap to be welded in the image are segmented using threshold selection. The boundaries of the threshold selection segment are the extreme values of the pixel distribution of the entire image in brightness. For unambiguous recognition of the desired edges based on the segmentation result, the features defining them as a pair of continuous, conditionally parallel lines are formalized. Relative to the recognized pair of edges, the correct position of the planned trajectory point relative to the current position of the welding head is determined. To transfer the correct position, we formalized the calculation model and chose an arbitrary point in the flat image in the workspace laser robotic welding complex, considering the orientation of the tool and the position of the camera. The results obtained made it possible to develop a correction system and successfully test it in the industrial complex.