Skip to main content

Research Repository

Advanced Search

Illumination compensation for nominally planar surface recovery

Sun, J.; Smith, M.; Smith, L.; Farooq, A.

Authors

J. Sun

M. Smith

L. Smith

A. Farooq



Abstract

A uniform distribution of illumination radiance across the surface of objects of interest is desired for most scene analysis approaches, including that of photometric stereo. Either specially designed parallel illumination or distant point light sources have been used to fulfil this requirement. However, there can be significant practical difficulties in achieving such idealised parallel light sources in real world applications, where setup space may be constrained and only generally available lights and lighting condition are provided. To solve this important but often overlooked problem, a flat reference object is proposed to sample the uneven incident lighting condition in order to allow the illumination radiance factor at every individual pixel to be cancelled out for the given scene analysis task. Experiments show that the proposed strategy can improve the accuracy of the photometric stereo method for reconstructing nominally planar surfaces containing topographic relief, even when normal low-cost commercial light sources are used. © 2012 The Institution of Engineering and Technology.

Citation

Sun, J., Smith, M., Smith, L., & Farooq, A. (2012). Illumination compensation for nominally planar surface recovery. IET Computer Vision, 6(5), 371-377. https://doi.org/10.1049/iet-cvi.2011.0075

Journal Article Type Article
Publication Date Sep 1, 2012
Journal IET Computer Vision
Print ISSN 1751-9632
Electronic ISSN 1751-9640
Publisher Institution of Engineering and Technology (IET)
Peer Reviewed Peer Reviewed
Volume 6
Issue 5
Pages 371-377
DOI https://doi.org/10.1049/iet-cvi.2011.0075
Keywords stereo image processing, image reconstruction, lighting, photometry, nominally planar surface reconstruction, nominally planar surface recovery, photometric stereo method, scene analysis task, flat reference object, scene analysis approach, parallel illum
Public URL https://uwe-repository.worktribe.com/output/955933
Publisher URL http://dx.doi.org/10.1049/iet-cvi.2011.0075

Downloadable Citations