A Study on Image Processing Algorithms for Seam Tracking System in GMA Welding

摘要

With the development of economy, the seam tracking technology for arc welding becomes one of the major research tasks in the manufacturing area with robots. In this study, the objectives aim to develop an intelligent and cost-effective algorithm based on the laser vision sensor for image processing in Gas Metal Arc (GMA) welding. Images of welded seam were captured from the CCD camera. These images were then processed by the algorithms in the proposed image processing. To optimize the effective image process, popular algorithms in use were verified, compared and finally selected for every step in the image processing. Moreover, owing to the simple interactive environment and abundant toolboxes, MATLAB was employed to realize those algorithms, which offer a sample for engineers to achieve the goal of algorithm developed by this new but easier approach. Finally, weld seam images obtained with different welding environments were processed to enhance the proposal validity, and it’s proved to have significant effect of getting rid of the variable noises to extract the feature points and centerline for seam tracking in GMA welding and is capable for industrial application.