Ellipsoid Criterion and Fuzzy C Means Algorithm for 3D Point Cloud Data Denoising

This paper describes a new method to extract most of the noise points by combining the ellipsoid criterion with the Fuzzy C-Means clustering algorithm. The point cloud data are unorganized and without any normal or orientation information. Firstly, we determine if one point is the noise or not by the ellipsoid criterion. After acquiring new point sets being less noisy, we delete large-scale noise points, and partly smooth small-scale noise with the Fuzzy C-Means clustering. The cluster centers are regarded as the new points. The experimental results show that the algorithm is a robust noise detection one.