The overall goals of my current research projects are to develop new methods of seabed mapping and to use those methods for ecological assessment of coral reef resources. Three complementary projects are underway, each working at a different spatial scale. One project maps the distribution of Acorpora palmata (elkhorn coral) using airborne hyperspectral imagery. A second project locates suitable grouper habitat using ship-based acoustic techniques. The third project uses video mosaicing technology to map individual plots (10-20 meters on a side) in very high detail (millimeter resolution).

As of Winter 2007-2008, highlights of the results of these projects include:

A) Airborne hyperspectral mapping

• Acropora palmata is spectrally distinct from other common Caribbean coral species, based on the spectra acquired in this project.
• Future hyperspectral image missions that seek to map A. palmata should place strong priority on very high spatial resolution (1 m or less) and accurate geolocation (better than one pixel).
• The results have been presented at four conferences. NASA scientists are using the spectral library produced from this work for underwater radiative transfer modeling.

B) Ship-based acoustic mapping

• An inexpensive single beam echo sounder was used to map substrate type and relief for the Navassa National Wildlife Refuge. NOAA scientists used the resulting map to conduct coral and reef fish resource assessments of the Refuge. The ability to generalize this new mapping technique to other areas is being assessed.
• The same system was used to characterize the site of a grouper aggregation in the Florida Keys. The seabed in that area was found to have higher slope, higher spatial variability, and closer proximity to deep offshore reefs relative to the surroundings. Other sites are now being surveyed to assess the generality of these results.
• Results have been presented at eight conferences. One publication is out and two are in press.

C) Diver mapping with video and multispectral imagery

• A technique for generating underwater video mosaics without external navigation has been developed. Mosaics covering about 500 m^2 with 1-2 mm spatial resolution are now routinely possible. Species identification is not always possible from the video, but higher order taxonomy (coral, algae, sponge etc..) is reliable.
• Spatial accuracy of the mosaics on the order of a few cm, which means that these tools are useful for rapidly measuring sizes and distances underwater. Applications include: coral demographics (size distribution), and measuring damages from hurricanes and ship groundings.
• The use of narrow spectral band underwater imagery enabled automated segmentation of coral and algae from the background and, in conjunction with a very simple texture classifier, enabled the classification of broad benthic cover categories: coral, algae, "other".
• Results have been presented at 7 conferences. Four publications are out or in press, and one paper is availbale from the Oceans 2007 conference, where it won the best student poster award.

The immediate application of these technological advances is coral reef ecology, but the mapping tools also could be used in other areas of ocean exploration or coastal resource management.

The following are sites that are known to link to this page. If you found me through one of these, click to take you back. If you found me through a web search, click to learn more about these organizations.
The University of Miami Rosenstiel School of Marine and Atmospheric Science (RSMAS).
The Division of Marine Geology and Geophysics at RSMAS.
The National Center for Coral Reef Research at RSMAS.
Home page for the underwater landscape video mosaic project.

A collection of experiences and example code that others might find useful.

My experience with the MATLAB Parallel Computing Toolbox

Last Updated: Feb 2008