OBJECTIVE
Utilize graphics processing units (GPUs) to accelerate an advanced hyperspectral remote sensing algorithm to near real-time analysis.
FUNDING AGENCY
U.S. National Science Foundation
PARTNER
Northeastern University
SOLUTION
This was the project that launched HySpeed Computing and put the company on the map. It was a combination of hyperspectral remote sensing, coastal habitat assessment, and high performance computing – a precursor of the many achievements that brought us where we are today.
Based on remote sensing analysis for mapping and monitoring of coral reefs, results illustrated that processing speeds for a compute-intensive hyperspectral optimization model could be improved to near real-time performance rates using a single GPU. As one of the earlier geospatial applications in the then emerging field of GPU computing, this innovative project demonstrated new capabilities for accelerating remote sensing data processing and analysis.
DATA
Hyperspectral | AVIRIS
PUBLICATIONS & PRESENTATIONS
Goodman J, Sellitto M, Kaeli D, 2012, Algorithm Acceleration for Geospatial Analysis presented at 2012 GPU Technology Conference, San Jose, California.
Sellitto M, Kaeli D, Goodman J, 2012, Accelerating an Imaging Spectroscopy Algorithm Using GPUs presented at GPU Technology Conference, San Jose, California.
Goodman J, Sellitto M, Kaeli D, 2012, Harnessing the Power of GPU Computing for Remote Sensing presented at VISualize 2012, Washington, DC.
Sellitto M, 2012, Accelerating an Imaging Spectroscopy Algorithm for Submerged Marine Environments Using Heterogeneous Computing Masters Thesis, Department of Electrical and Computer Engineering, Northeastern University.
Goodman J, 2011, Developing a Remote Sensing Analysis Module for Shallow Marine Environments presented at VISualize 2011, Washington, DC.
Image Data (top, bottom): AVIRIS/JPL/NASA
