DocumentCode :
3315586
Title :
Pixel mapping for generation of 3D coloured seafloor bathymetry using a single camera
Author :
Bodenmann, Adrian ; Thornton, Blair ; Ura, Tamaki ; Sangekar, M.
Author_Institution :
Underwater Technol. Res. Center, Univ. of Tokyo, Tokyo, Japan
fYear :
2010
fDate :
1-3 Sept. 2010
Firstpage :
1
Lastpage :
9
Abstract :
Visual surveys of the seafloor are often performed by robotic underwater vehicles and a number of methods exists for transforming photos or video into 2D or 3D reconstructions, most of which utilise navigation data. Laser profiling is a method that is widely used for bathymetry mapping with high vertical resolution, but typically no information of the actual colour is obtained. In this paper we propose a method for high resolution three dimensional mapping of the seafloor in colour based on laser profiling and retrieval of colour information through pixel based mapping. It employs a compact algorithm treating the data in a fully three dimensional way from input to output, which does not yield visual artifacts usually encountered when employing methods assuming planar scenes. The same hardware as for laser profiling (one video camera, a sheet laser and a navigation sensor), plus a lamp are sufficient for recording the required data. The sensors require only little space, allowing the system to be used in missions where other data is collected as the main task. Combining the 3D map with that data would allow the information to be plotted as a geographic information system (GIS) in the future.
Keywords :
bathymetry; geographic information systems; geophysical image processing; image colour analysis; 3D coloured seafloor bathymetry; GIS; bathymetry mapping; geographic information system; high resolution three dimensional mapping; laser profiling; pixel mapping; single camera; Cameras; Image color analysis; Image resolution; Lasers; Pixel; Three dimensional displays; Vehicles;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Autonomous Underwater Vehicles (AUV), 2010 IEEE/OES
Conference_Location :
Monterey, CA
ISSN :
1522-3167
Print_ISBN :
978-1-61284-980-5
Type :
conf
DOI :
10.1109/AUV.2010.5779673
Filename :
5779673
Link To Document :
بازگشت