Engineering student fuses together satellite data to aid in emergency response
Benjamin Martin, an engineering student at the Univ. of Tennessee enrolled in the Higher Education Research Experiences program at Oak Ridge National Laboratory, is trying to improve emergency response by developing computer codes to fuse together two different kinds of satellite image data. In the event of a man-made or natural disaster, Martin’s codes could be used to enable a comparison of pre- and post-disaster images to determine areas with the highest level of damage.
A funnel cloud falls from the sky and twists into a tornado, tossing trucks and plucking up pines. As the rain floods the creeks with runoff, people crawl out from their hiding spaces and begin to assess damage. With phone lines down, emergency crews can be slow to determine which areas are of highest-impact and need.
As part of the Higher Education Research Experience, or HERE, program, University of Tennessee (UT) engineering student Benjamin Martin conducts research at Oak Ridge National Laboratory (ORNL) that could aid in post-catastrophe emergency response. The HERE program, administered by the Oak Ridge Institute for Science and Education, managed by ORAU, provides research opportunities for students and faculty and is designed to complement academic programs by utilizing ORNL resources to enhance science, technology, engineering and math (STEM) education.
“The idea is to detect changes in terrain,” Martin said, describing how his current project seeks to integrate both optical and synthetic aperture radar, or SAR, data to determine areas of concern—even on a global scale. “If you have a whole region that has been damaged by a catastrophe, you want to be able to analyze it quickly by identifying key changes, such as which infrastructures are down or what buildings are damaged.”
Senior Research Scientist Raju Vatsavai, Martin’s mentor, said human and economic losses due to major natural and man-made disasters are increasing.
“This increase can be attributed to growing population, human migration, and settlements in disaster-prone regions of the world,” he said. “Rapid damage assessment and dissemination of accurate information is critical for creating an effective emergency response.”
The basis for Martin’s research project draws from the differences between optical and SAR data acquired before and after the event.
“The benefit of using the two types of data together is that the optical data can provide a high level of detail whenever there is not substantial cloud cover or related weather issues,” said Martin, “The radar data can be used to fill in the gaps during times when sufficient optical data cannot be collected for whatever reason.”
He explains that the SAR data is harder to acquire and interpret due to noise in the radar signal, but it can prove helpful in certain situations, such as fires and hurricanes. “Any optical data collected during these events could consist mostly of cloud and smoke images and may not be useful when trying to assess damage or map terrain for rescue workers.”
To carry out his project, Martin researches and develops algorithms to fuse together optical and SAR data images in order to generate change and damage maps.
“Data fusion is more than merely overlaying two images,” he explained. “It attempts to take the most important information from either image and combine it mathematically into one map; the combination of this data will be helpful to emergency responders, because the fused data can be compared to the SAR data when the optical data is not available.”
He spends most of his time programming on a desktop supercomputer, using C/C++, Java, Python, MATLAB and CUDA computer programing languages.
Martin applies these programming skills not only to his appointment at ORNL, but also to his membership in UT’s Institute of Electrical and Electronics Engineers (IEEE) Graduate Robotics Team. In the IEEE SoutheastCon 2013 in Jacksonville, Fla., Martin and his teammates placed first ahead of eight other university teams in the Open Hardware Division. Now, he is making plans to design a robot for next year’s competition. He said he envisions transferring some of the lessons learned from his robotics teams to his internship work at ORNL.
“Even though I was mostly working on control code for the competition, I did learn a lot about imaging concepts. I would imagine that some of what I learned could potentially be applied to the satellite research that I do as part of my HERE appointment,” he explained.
In addition to programming, Martin has also gained all-around research skills in his HERE appointment that he feels has made him a better researcher.
“The whole program at ORNL is just great; I cannot single out one thing that is better than another,” he said. “By far, participating in the HERE program definitely has been the best thing I have done academically and career-wise.”