文摘
A method is introduced to investigate the interrelationships among welding penetration rate, spatter number, metallic vapor, and welding quality by simultaneously observing both top surface and bottom surface of low carbon steel weldment during high-power disk laser welding. Color image segmentation algorithm which is based on K-means clustering algorithm is used to process the image data. Different welding conditions including different weld power, weld speed and weld joint width are applied, and the microstructures of fusion part are also analyzed. Experiment results show that the color image segmentation algorithm is effectively for recognition of welding penetration condition, and welding quality is better when the penetration rate is around 74.5 %.