An on-line shape-matching weld seam tracking system

• A seam tracking system with high adaptability to groove types was developed.
• The system is able to localize the starting/finishing point of the weld seam.
• A FIFO based queue was implemented to tackle the lag distance problem.
• The system was verified on a sine-shaped seam with an accuracy of ±0.5 mm.

Automatic welding technology continues to find a broader application in diverse industries due to its high efficiency and accuracy. In this work, an on-line laser-based machine vision system for seam tracking was developed. To achieve a reliable and accurate seam tracking process that is adaptive to different groove types, a shape-matching algorithm was proposed and implemented. The algorithm uses the previous groove shape as the template to locate the next groove shape. Tests on U-groove, tap-groove and free form groove have verified its adaptability and robustness to different groove types with noise. The shape-matching algorithm also enables the seam tracking system to automatically localize the starting and finishing point of the weld seam. A FIFO based queue was defined and implemented to tackle the lag distance problem between the heat source and the vision sensor. The tracking algorithm, along with the FIFO queue was successfully verified on a sine-shaped seam. A tracking accuracy of ±0.5 mm was achieved in this test, which is acceptable in most of the arc welding applications.