3D Point Smoothening Using Modified Local Regression for Reverse Engineering Process

3rd International Conference on Mechanical, Automotive and Aerospace Engineering 2016

机械制造;无线电电子学;航空航天工程

Abdulqawi, Nur Ilham Aminullah^1 ; Abu Mansor, Mohd Salman^1

School of Mechanical Engineering, Engineering Campus, Universiti Sains Malaysia, Nibong Tebal, Pulau Pinang

14300, Malaysia^1

关键词: Curve-fitting tool;    Local regression;    Non-smooth surface;    Reverse engineering process;    Statistical datas;    Statistical tools;    Surface conversions;    Surface details;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/184/1/012052/pdf
DOI  :  10.1088/1757-899X/184/1/012052

学科分类：航空航天科学

来源: IOP

【 摘 要 】

This research will introduce an existing statistical curve fitting tool i.e. local regression into reverse engineering process. Instead of using the local regression on statistical data, it will be applied on 3D point clouds obtained from 3D scanners. As most 3D scanners capture the smallest surface details such as textured and non-smooth surfaces, it makes the 3D point cloud to surface conversion process problematic. By modifying this existing statistical tool of local regression with the proposed algorithm, it helps to smooth out rough surfaces obtained from 3D point cloud, minimising the error when generating the virtual surface from smoothened point cloud rather than from raw point cloud. The related method of performing the smoothening process on the 3D point will also be included in this paper. At the end, comparison of the smoothened point cloud and the unsmoothened point cloud will be visualised and discussed.