Weld seam tracking deviation extraction based on multi-sensor information fusion

Objective The aim is to study the mechanism of extracting weld seam tracking deviation and the method of improving tracking accuracy based on the fusion of binocular vision sensors and laser displacement sensors. Methods By combining theory and simulation, this study aims to address the problem of low accuracy in weld seam tracking caused by multiple interference factors, low reliability and fault tolerance of single sensor information acquisition during automatic weld seam tracking. Based on the fusion mechanism of two types of sensors, this study explores ways to improve the deviation extraction method for weld seam tracking based on the fusion of binocular vision sensors and laser displacement sensors. Firstly, position information of weld seam is obtained through binocular vision sensors and laser displacement sensors respectively. Then, based on Hough transform, key position feature points are extracted from weld seam image. Finally, based on the improved adaptive weighting method, weld seam tracking information obtained from two types of sensors is fused to obtain the final weld seam tracking deviation information. Results The research results indicate that welding seam tracking deviation based on information fusion method of binocular vision sensors and laser displacement sensors can be controlled within a reasonable range. Conclusion Therefore, sensor weld seam information fusion method based on Hough transform and adaptive weighting can effectively improve information acquisition accuracy of weld seam tracking.