Algorithm for 3D Point Cloud Denoising

The raw data of point cloud produced by 3D scanning tools contains additive noise from various sources. This paper proposes a method for 3D unorganized point cloud denoising by making full use of the depth information of unorganized points and space analytic geometry theory, applying over-domain average method for 2D image of image denoising theory to 3D point data. The point cloud noises are filtered by using irregular polyhedron based on the limited local neighborhoods. The experiment shows that the proposed method successfully removes noise from point cloud with the features of the scattered point model reserved. Furthermore, the presented algorithm excels in its simplicity both in implementation and operation.