An efficient partitioning method for curved-surface laser texturing based on bipolar projection and multi-level pruning

Abstract

Laser surface texturing (LST), as a highly promising technology, is applied to surface functional enhancement by forming special micron/nanometer-level texture structures. Curved-surface LST requires surface partitioning to satisfy laser process constraints. Dedicated partitioning methods for this application remain limited. This paper proposes a Freeform Mesh Partitioning Method Based on Bipolar Projection and Multi-Level Manhattan-Euclidean Pruning. Bipolar projection is employed for rapid angular candidate mapping, and multi-level Manhattan-Euclidean pruning is utilized for progressive pruning, thereby enabling the efficient and regular partitioning of complex freeform surfaces. Tests on Armadillo, Typical Surface, and Bunny models under 12 orthogonal cases show reductions of 69.5%/39.6% in partition number, 21.8%/16.9% in boundary length, and 60.0%/93.0% in computation time versus Watershed and Random Seed, while producing more complete regions and regular boundaries. Consequently, an effective solution for efficient partition planning in curved-surface LST is provided.