The Office of Polar Programs, Antarctic Science Division, Ocean & Climate Systems Program has made this award to support application of digital cameras in the visual and near infra-red wavelengths with accompanying development of software to stereoscopically and quantitatively capture what human vision has observed for decades regarding sea ice distributions and thickness in the polar regions. The overarching goal is to improve understanding of sea ice thickness variability by quantitatively enhancing ship-based sea ice observations typically done by humans. Advances in digital photography and computer technology, speed, and data storage over the past five years now make it feasible to apply techniques from the field of computer vision to the determination of sea ice surface topography, roughness, and feature characterization. The investigators will create software system for rendering three-dimensional sea ice topography to a maximum distance of 100 m obliquely from the side of a moving ship. A sequence of four-camera (two visual and two near infra-red) images will be reduced to a composite map of estimated x, y, z values. The composite field will have an associated uncertainty map providing information about the precision of each data point and a data header containing meta information including number of images, camera specifications, and global positioning information. The system will undergo a development phase to include deployment in a river ice setting and then be deployed from a ship in Antarctica as a further test of concept.
Broader Impacts
This technology will provide swath-type information at the local scale that can be accumulated and applied to satellite algorithm and model validation. It has the potential to provide an economical and effective supplement to LIDAR systems, ice core based determinations, and moored or autonomous vehicle deployed upward looking sonar, which are currently used to determine sea ice thickness, an essential but until now more elusive sea ice variable than surface area. This project will support a Ph.D. student, be used in new college course development and develop a product that can be integrated into the existing NSF-supported 3D GIS visualization tool called EarthSlot (www.earthslot.org/).
