Fabrication accuracy evaluation of ship spatial pipes using NURBS-based surface reconstruction from point clouds

Geometric deviations in the fabrication of ship spatial pipes caused by cutting inaccuracies, welding-induced distortion, and bending errors pose challenges to manufacturing quality control. Manual inspection remains limited by operator subjectivity and sparse discrete measurements, which cannot capture the complete 3D morphology of complex piping geometries. This paper proposes an automated point-cloud-based surface reconstruction framework for accuracy evaluation of ship spatial pipes. First, a skeletonization algorithm integrated with center correction is developed to robustly extract pipe centerlines. Next, an adaptive cross-sectional slicing strategy is applied along the extracted skeleton to construct a topologically consistent control-point matrix from unorganized point clouds. Finally, a high-fidelity Non-Uniform Rational B-Spline (NURBS) surface is fitted to these control points, enabling systematic quantification of geometric discrepancies. Experiments on three typical pipe specimens demonstrate the effectiveness and robustness of the proposed method, indicating its potential to support digital quality assessment of ship spatial pipes.