Share Email Print

Optical Engineering

Automated detection of planes in 3-D point clouds using fast Hough transforms
Author(s): Olatokunbo O. Ogundana; Jonathan M. Huntley; C. Russell Coggrave; Richard L. Burguete
Format Member Price Non-Member Price
PDF $20.00 $25.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Calibration of 3-D optical sensors often involves the use of calibration artifacts consisting of geometric features, such as 2 or more planes or spheres of known separation. In order to reduce data processing time and minimize user input during calibration, the respective features of the calibration artifact need to be automatically detected and labeled from the measured point clouds. The Hough transform (HT), which is a well-known method for line detection based on foot-of-normal parameterization, has been extended to plane detection in 3-D space. However, the typically sparse intermediate 3-D Hough accumulator space leads to excessive memory storage requirements. A 3-D HT method based on voting in an optimized sparse 3-D matrix model and efficient peak detection in Hough space is described. An alternative 1-D HT is also investigated for rapid detection of nominally parallel planes. Examples of the performance of these methods using simulated and experimental shape data are presented.

Paper Details

Date Published: 1 May 2011
PDF: 12 pages
Opt. Eng. 50(5) 053609 doi: 10.1117/1.3562323
Published in: Optical Engineering Volume 50, Issue 5
Show Author Affiliations
Olatokunbo O. Ogundana, Loughborough Univ. (United Kingdom)
Jonathan M. Huntley, Loughborough Univ. (United Kingdom)
C. Russell Coggrave, Phase Vision Ltd. (United Kingdom)
Richard L. Burguete, Airbus Operations Ltd. (United Kingdom)

© SPIE. Terms of Use
Back to Top