UofL engineer reduces land mine dangers
November 26th, 2007
By Kevin Rayburn
Hichem Frigui
In the time it takes for a typical American to commute to work or to take an evening jog, two more people somewhere in the world have lost a limb, their sight or their lives to land mines.
That’s because an estimated 60 million to 100 million land mines or other unexploded weaponry lie buried in more than 70 countries — even peaceful ones. Often, the victims of these remnants of past or ongoing wars are unintended targets: civilians.
New mine detectors developed by University of Louisville researcher Hichem Frigui, an associate professor of computer engineering and computer science in the J.B. Speed School of Engineering, are helping to find and clean old and new land mine sites.
Although traditional detectors can locate metal mines with relative ease, mines and other explosive devices increasingly are made with plastics and other hard-to-detect materials. Some are buried just a few inches from the surface while others can lie 2 feet or more below the surface.
In the battlefield, detection can bog down an army that needs to move quickly. That’s why the U.S. Department of Defense increasingly is tapping the expertise of engineers at American universities.
With financial support from the Defense Department, Frigui works in tandem with research colleagues at other universities and with companies such as Niitek and BAE to improve land mine detection.
One of his goals is to improve the way detectors interpret data gathered from the ground-penetrating radar used on special mine-sweeping military vehicles and hand-held devices.
“An army can’t stop for every object,” Frigui said. “You want to know if that suspicious object is a real mine or just another tin can. Sometimes, with older methods of detection, it can be hard to tell the difference.”
Challenges to accurate detection include the wide variety of land mine types, shapes, sizes and materials. Detector effectiveness also can vary according to terrain or soil type — even water or snow on the ground can affect data readings.
To tackle these problems, Frigui and his research team are developing algorithms — computer-based mathematical formulas — that can help detectors better filter out extraneous visual “noise” and present a clearer picture of buried objects.
Over a four-year period, Frigui has used his team’s algorithms and related mathematical models at military bases across the country to analyze data from all sorts of mine types buried in simulated minefields under grass, dirt, asphalt, snowy ground and other surfaces.
One of his goals has been to reduce the number of “false alarms.”
“We’ve been able to do that,” Frigui said. “In the best-case scenario, we can get roughly 95 percent (accurate) detection.”
The improved performance is the result of the fusion of four different algorithms, some developed at UofL and some developed elsewhere.
“It’s a great example of cooperative research,” Frigui said.
So far, the U.S. Army has been happy with the results. Although Frigui and the Army won’t say where the newest detectors that incorporate his improvements are being used, they will say that they are helping to protect U.S. soldiers.
“It’s rewarding to know that the research you do is not only being used right away in the field but is maybe even saving lives,” Frigui said.
