Weld seam location with adaptive fracture repair on minor-caliber pipe

Le résumé fourni par la source

High-precision weld positioning is crucial for enhancing welding quality and structural reliability. Current algorithms are challenged by uneven laser reflectivity, complex surface curvature, and noise interference. To overcome these obstacles, we propose a multi-stage adaptive weld centering framework integrating line laser scanning, skeleton generation and tracking, adaptive fracture repair techniques. The framework consists of three primary stages: (1) Image preprocessing utilizes a hybrid median filter and fixed threshold segmentation to enhance contrast, suppress salt-and-pepper noise, and preserve edge details. (2) Centerline extraction employs an improved maximal disk-based skeleton algorithm to generate single-pixel-width centerlines, followed by Freeman chain code tracking and data storage. Connectivity analysis via BFS ensures topological consistency among fragment skeletons. (3) Adaptive repair and smoothing method adopts least squares optimization to integrate broken skeleton segments through geometric constraints, achieving continuous sub-pixel trajectories. Positioning experiment of pipeline welds validates the algorithm effectiveness by a maximum positional error of 3.6 pixels and an average error of 0.98 pixel. The deviation rate of the weld center point from the manually calibrated 40.68 pixels is below 5%. These results meet the accuracy standards necessary for practical welding applications.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.