High-resolution Depth and Reflectance Map Estimation in Underwater Imaging

Laszlo Neumann, Rafael Garcia, Jordi Ferrer, Attila Neumann
High-resolution Depth and Reflectance Map Estimation in Underwater Imaging
TR-186-2-04-14, December 2004
Content:

Information

Abstract

This paper describes a novel technique to simultaneously determining the threedimensional geometry and reflectance map of a scene in underwater imaging. Underwater light is affected by an exponential wavelenght-dependent absorption and undesired scattering effects. These special conditions make image and video processing a difficult task in participating media. It is shown how to use this disturbing absorption effect to find a new way to accurately estimate the scene depth-map and multispectral reflectivities. The process requires the acquisition of a multispectral image of the underwater scenario with the appropriate camera and lightsource arrangement. The algorithm has been tested on several synthetic images for different noise levels. The promising results show that the technique provides accurate and stable results.

Additional Files and Images

No additional files or images.

BibTeX

Download BibTeX-Entry
@techreport{Neum04Hig,
  title =      "High-resolution Depth and Reflectance Map Estimation in
               Underwater Imaging",
  author =     "Laszlo Neumann and Rafael Garcia and Jordi Ferrer and Attila
               Neumann",
  year =       "2004",
  abstract =   "This paper describes  a novel technique to simultaneously
               determining the threedimensional geometry and reflectance
               map of a scene in underwater imaging. Underwater light is
               affected by an exponential wavelenght-dependent absorption
               and undesired scattering effects. These special conditions
               make image and video processing a difficult task in
               participating media. It is shown how to use this disturbing
               absorption effect to find a new way to accurately estimate
               the scene depth-map and multispectral reflectivities. The
               process requires the acquisition of a multispectral image of
               the underwater scenario with the appropriate camera and
               lightsource arrangement. The algorithm has been tested on
               several synthetic images for different noise levels. The
               promising results show that the technique provides accurate
               and stable results.",
  address =    "Favoritenstrasse 9-11/186, A-1040 Vienna, Austria",
  institution = "Institute of Computer Graphics and Algorithms, Vienna
               University of Technology",
  note =       "human contact: technical-report@cg.tuwien.ac.at",
  month =      dec,
  number =     "TR-186-2-04-14",
  keywords =   "Underwater Imaging, Depth Map",
  URL =        "http://www.cg.tuwien.ac.at/research/publications/2004/Neum04Hig/",
}