Laser line detection for groove butt joint type based on vision system

• Published in: AIP conference proceedings Vol. 2795; no. 1
• Main Authors: Johan, Nurul Fatiha, Shah, Hairol Nizam Mohd, Sulaiman, Marizan, Naji, Osama Abdullah Ahmed Mohammed
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 25.05.2023
• Subjects: Butt joints; Grooves; Hough transformation; Image processing; Laser beam welding; Lasers; Metal surfaces; Seam tracking; Seam welds; Shape; Vision systems; Workstations
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0122614

An extraction of weld seam groove image gave a huge impact in weld seam tracking to produce a quality welding system. The big challenge to deal with vision and image processing normally comes from lighting, workstation environment uncertainty, reflection from a material surface, or metal surface imperfections. Differences in either laser images or work environments may require different methods to achieve an outperformed result. Hough transform is a dominant method in the field of line detection where it can detect all possible lines that exist in an image but sometimes did not suitable for a particular application. This paper introduced a new approach of line detection based on pixel scanning and reference row index as a guide to track the weld seam position line and groove line from the V-shape image. The results showed that the proposed line detection is more applicable to be used compared to the Hough transform especially if the lines are uneven along the laser skeleton line and the same time can be used to fix the laser line image.