Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Improving Fitting CAD to 3D Point Cloud Acquired with Line-of-Sight Sensor

Published

Author(s)

Marek Franaszek, Prem Rachakonda, Kamel S. Saidi

Abstract

Due to self-occlusions in 3D point clouds acquired with line-of-sight sensors, an incomplete representation of an object's surface is used in fitting a CAD (computer-aided design) model to the data. For certain types of object geometry, CAD pose fitted using the Iterative Closest Point (ICP) procedure is systematically misaligned, even when the selected initial pose is very close to the expected, correct pose. We demonstrate on experimental data obtained from manufacturing-relevant parts that the final CAD alignment can be greatly improved if only a part of the CAD surface is used in the ICP registration. The resulting residual ICP error is then reduced by three to four times. We also introduce another measure which can be used to gauge pose alignment and show that both metrics are well correlated most of the time. In rare occasions, a large discrepancy between the two metrics is observed, and in these cases the new measure provides a better gauge of pose alignment.
Proceedings Title
2023 IEEE International Symposium on Robotic and Sensors Environments (ROSE)
Conference Dates
November 6-7, 2023
Conference Location
Tokyo, JP

Keywords

6DOF pose, 3D point cloud, ICP, bin picking

Citation

Franaszek, M. , Rachakonda, P. and Saidi, K. (2023), Improving Fitting CAD to 3D Point Cloud Acquired with Line-of-Sight Sensor, 2023 IEEE International Symposium on Robotic and Sensors Environments (ROSE), Tokyo, JP, [online], https://doi.org/10.1109/ROSE60297.2023.10410732, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=936471 (Accessed April 23, 2024)
Created November 7, 2023, Updated March 5, 2024