Landmine detection isn't easy. Everything from rubbish to rabbits can cause a false alarm. Maths PhD student John Schofield has been working on algorithms to make clearing minefields safer.
Between 45 and 100 million unexploded land mines lie beneath the ground today, spread over 75 countries. Over the decades, injuries and deaths from antipersonnel land mines are thought to run into hundreds of thousands. Land mines have been used in many conflicts over the years, and just one of the complex issues for recovering countries is their detection and safe removal.
Our Maths researchers have developed a new technique that substantially improves the detection of land mines. Antipersonnel land mines are small, explosive devices designed to injure or kill. Hidden in or on the ground, they are triggered by pressure and kill indiscriminately: soldiers, peacekeepers, aid workers, and children.
The mines can lie dormant for decades, long after conflicts have ceased and the only way to deactivate them is by individual removal, posing grave risk to trained disposal experts. Detection of non-metallic mines typically uses ground-penetrating radar (GPR), sending electromagnetic waves deep into the ground to detect any reflected signal caused by buried objects.
The ground disperses the electromagnetic signal and this is affected by natural changes in soil density or moisture. The challenge in GPR detection is in pin-pointing the location of the potential mine without knowing the exact properties of the ground. Submerged items such as plant roots or rocks can also cause false alarms, making detection a time-consuming and costly process.
PhD student John Schofield, working with Dr Paul Hammerton, developed a new image ‘stacking' algorithm, which greatly reduces the effect of fluctuations in ground conditions on GPS signals. Results for a range of diffraction parameters are combined in such a way that the uncertainty in the location of the target is minimised. Researchers created a test site where they buried replica mines in various ground types and tested how well the mathematical model coped with varying conditions.
The research was carried out as an internship with a cutting edge technology manufacturer, part of the KTN’s Industrial Mathematics KTP Programme which aimed to solve business problems by working in partnership with leading UK Universities.
John’s research is influencing the development of new mine detection technologies and contributing to global recovery from conflict.
John Schofield & Dr. Paul Hammerton