Analysis of weld seam characteristic parameters identification for laser welding of dissimilar materials based on image segmentation

Abstract

Welded joints of dissimilar materials increase the flexibility in design and manufacturing process greatly and, hence, have been widely used in aerospace, rail transportation, and other related industries. Due to the difference in physical and chemical properties of dissimilar materials, the formed weld during laser welding is different from that of the same material welding. The geometric morphology of the formed weld is an important factor affecting the welded joints performance. Therefore, an identification method of weld seam characteristic parameters is proposed in this paper for evaluating the welding quality by image segmentation in the laser welding of low carbon steel (Q235) and stainless steel (316L). The region of interest of the weld metallograph from experimental observation is defined, converted into grayscale image and then denoised by filter. The weld is segmented by the seeded region growing method with initial seed automatic selection. The weld seam characteristic parameters including the weld area, left weld width, right weld width, and weld penetration in the laser welding of dissimilar materials are identified based on the segmented image. The obtained results are validated by the experimental measurements of weld and good agreement between them has been found. The identified weld seam characteristic parameters are adopted for assessing weld fusion status, depth-to-width ratio, and symmetry quantitatively. The results indicate that the proposed method is reasonable and feasible for the weld quality evaluation to improve the laser welding quality of dissimilar materials in practical production.