Groove depth measurement based on laser extraction and vision system

• Main Authors: Mohd Shah, Hairol Nizam, Johan, Nurul Fatiha, Sulaiman, Marizan, Mohammed Naji, Osamah Abdullah Ahmed, Arshad, Mohd Ali
• Format: Article
• Language: English
• Published: Springer Science and Business Media Deutschland GmbH 2024
• Online Access: http://eprints.utem.edu.my/id/eprint/28341/2/01275241220241314131449.pdf; http://eprints.utem.edu.my/id/eprint/28341/; https://link.springer.com/article/10.1007/s00170-023-12914-9#:~:text=The%20vision%20sensors%20cooperate%20with,height%20of%20the%20groove%20joint.

Laser is a very useful technology in the field of welding to obtain the deepest point of the metal being joined by providing a light source at the weld seam. However, laser imaging may be challenging because there are various laser reflection profiles on the workpiece that make it difficult to extract the desired laser image. In this paper, an extension study on feature point extraction was proposed to determine the depth of the V-groove. By taking use of laser image which has intensity noise around the edges, a noise rejection technique is applied to improve the quality of laser image. A data fitting method for the purpose of extracting feature points based on the reference row has been proposed because it is suitable for use after the laser centerline search. Then, the feature points are obtained by assigning the orientation position based on the “V” shape. Afterward, the extension study from feature point extraction looks on how to determine the V-groove depth by using the intersection and distance measurement method. To validate the accuracy of the proposed method, several samples of straight line and half-moon line types were tested. The performance of the system was evaluated against the actual value. It was observed that the proposed method is acceptable when subjected to laser reflection and lighting variations. The proposed method matches accurately when the respective line types showed an error percentage within 2 to 6%. This study is an extension step from feature point extraction, which can provide further analysis for weld seam tracking applications.