A Delaunay Triangulation-Based Point Cloud Hole Filling Algorithm by Fusion of 2D and 3D Data

In complex industrial environments, LiDAR scanning encounters challenges such as environmental noise, high temperatures, dust, and metal reflections, often resulting in the occurrence of holes in point cloud data. This study presents a solution that begins with computing the projected points of the point cloud on a two-dimensional plane and generating a grid of subdivided arrays to delineate hole boundaries. Subsequently, the Delaunay algorithm is utilized to construct triangular meshes for hole filling and reconstructing missing surfaces. Experimental results confirm the method’s effectiveness in accurately identifying and filling hole boundaries.