Matching Planes as a Constellation

Report No. ARL-TN-0468
Authors: Gary Haas
Date/Pages: January 2012; 32 pages
Abstract: A robot exploring and mapping an unknown indoor environment may be equipped with a sensor which detects clouds of points and extracts from them planar structures such as walls. A model of the environment will be built from these structures, and the model will be confirmed and extended by point clouds subsequently sensed. This report describes and interprets two published methods for matching planes from a point cloud to corresponding planes from another point cloud, or from a model. Matches are based on intrinsic plane elements, and on the relative orientation of the planes one to another. One approach is compiled from a set of recent papers by Kaustubh Pathak and colleagues at Jacobs University, Bremen; the other is from a 2003 paper by Daniel Fischer and Peter Kohlhepp. The second approach is illustrated with a simplified example. This report is intended as a guide to the original papers discussed.
Distribution: Approved for public release
  Download Report ( 1.558 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.

Last Update / Reviewed: January 1, 2012