Normal Vector Estimation for Point Clouds via Local Delaunay Triangle Mesh Matching

• Reliable estimation of normal vectors for point clouds is of practical importance in computer-aided geometric modeling and inspection. This paper introduces a new normal vector estimation method for point clouds based on the matching results of the local Delaunay triangle mesh formed at each point. The local mesh is a manifold patch of Delaunay triangles and resembles an open umbrella. According to the matching results of these umbrellas, the local Delaunay neighbors at each point can be reliably identified, which leads to accurate normal vector calculations. Compared with the existing methods, the proposed method yields normal vectors of notably improved accuracy, especially for points near edge or corner features. The improvement has been demonstrated using both simulated and scanned point cloud data sets.

Publication and Content Type

ref (subject category)

art (subject category)