Image analysis techniques for marine and Antarctic applications
Dr. Ray Williams, Ph.D., University of Tasmania
Wednesday, August 9, 2006
Pacific Forum – 3:00 p.m.
SPOT Image of Heard Island, draped on a Digital Elevation Model
Provided by the UTAS Centre for Spatial Information Systems
The School of Computing, at the University of Tasmania, is currently expanding its Information and Communications Technology research program within marine and Antarctic science. This research involves the application of computer vision and machine learning techniques, combined with scientific visualization tools, to the automated/semi-automated analysis of images and video. One major objective is to detect, identify and track various marine animals, including scallops, lobster, octopus and krill in underwater video footage taken by fixed and towed cameras, with the possibility of using these techniques for future work on ROVs and AUVs. A second project is investigating the feasibility of using texture analysis of satellite imagery to classify vegetation types on sub-Antarctic islands. Further projects involve iceberg detection and monitoring off the Antarctic coast using synthetic-aperture-radar satellite images, analysis of multi-beam sonar data for fisheries applications and content-based retrieval of photographic images within the Antarctic Data Centre image repository.