UCT CS Research Document Archive

MLS reconstruction from noisy point sets

Merry, Bruce and Christoph Held (2011) MLS reconstruction from noisy point sets. Technical Report CS11-04-00, Department of Computer Science, University of Cape Town.

Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

For digital preservation of cultural heritage sites in Africa, laser range scanning has been used to produce point clouds. The literature contains extensive work on reconstructing surface models from such point clouds, but often this prior work does not account for artefacts in the data such as vegetation. We have assessed several variations on a specific moving-least-squares (MLS) technique to determine the impact on the quality of the reconstructed surfaces. We found that correct feature size detection and explicit detection of boundaries is important, while a single iteration of almost orthogonal projection is sufficient to give good results.

EPrint Type:Departmental Technical Report
Keywords:Moving Least Squares, Surface Reconstruction, Laser Range Scanning
Subjects:I Computing Methodologies: I.3 COMPUTER GRAPHICS
J Computer Applications: J.7 COMPUTERS IN OTHER SYSTEMS
ID Code:737
Deposited By:Merry, Bruce
Deposited On:05 December 2011