A rapid image processing algorithm for structured light stripe welding seam tracking in laser welding

Welding seam tracking plays an important role in improving welding quality, reducing rejection rate and repair rate in which accurately and real-time image processing is one of the most important objectives. Butt weld without groove in laser welding increases the difficulty of seam detection. A seam tracking system for laser welding was established based on structured light strip and CCD camera. Image from CCD are processed twice for better weld-recognition rate, in which one uses high-pass filter for structured light stripe line, and the other low-pass filter for weld seam curve. Binarization threshold processing, sobel edge detection and gravity center algorithms were employed in the system. Weld gun, then, can be adjusted in horizontal and vertical direction according to the intersection of the stripe line and the weld seam curve. Preliminary test results show that the algorithm improves weld-recognition rate and lowers image processing period for welding seam tracking.