Using artificial intelligence, a team of researchers calculated the number of possible unexploded bombs from the Vietnam War. The new method has increased the detection of over 160% compared to standard methods. The latter also suggests that 44-50% of the bombs in the studied area could be unexploded.
There have been several attempts to safely find and remove unexploded bombs and anti-personnel mines, but they have not been as effective as needed in Cambodia, says Erin Lin of Ohio State University. So the assistant professor co-conducted a study with Rongjun Qin, a colleague from his university. The researchers have started with a commercial satellite image of an area of 100 square kilometres near the city of Kampong Trabaek, an area that was heavily bombed by the U.S. Air Force from May 1970 to August 1973.
Researchers know how many bombs were dropped in the area and the general location of where they fell. The craters of the explosion tell them how many bombs went off and where. They can, therefore, determine how many unexploded bombs are left and the specific areas where they could be found. To calculate the whole, artificial intelligence was used. After AI learned to detect the real craters created by the bombs, one of the researchers checked the computer’s work. In the end, the algorithm achieved an accuracy rate of approximately 86%, identifying 152 of 177 craters.
Much of the territory covered by this study is agricultural, which means that Local farmers are at risk of coming across an unexploded bomb. 60 years after the bombing of Cambodia, over 64,000 people were killed or injured by unexploded bombs. Currently, the injury count is on average one person each week.