Robust recovery of 3D geometric primitives from point cloud.

Xiang Yang(1), Peter Meer(2) and H.G. Gea
(1)Dept. of Mechanical and Aerospace Engineering
(2) Dept. of Electrical and Computer Engineering
Rutgers University, NJ 08854, USA

A robust method for surface fitting in 3D point cloud is presented as an application of the robust estimation of multiple inlier structures algorithm [1]. The geometric primitives such as planes, spheres and cylinders are detected from the point samples in the noisy dataset, without regenerating surface normals or mesh. The inlier points of different surfaces are classified and segmented, with the tolerance of error for each surface estimated adaptively from the input data. From the segmented points, designers can interact with the geometric primitives conveniently. Direct modification of 3D point cloud and inverse design of solid model can be applied. Both synthetic and real point cloud datasets are tested for the use of the robust algorithm.
ASME 2017 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference, 37th Computers and Information in Engineering Conference, CIE-8 Computer-Aided Product and Process Development. Cleveland, Ohio, August 2017, DETC2017-67564.
Return to Research: Robust Analysis of Visual Data        Return to List of Publications
Download the paper