The International Committee of the Red Cross seeks to utilize Japan-developed artificial intelligence technology to speed up the detection of landmines and dud munitions in war-torn areas including Ukraine so displaced people can return home sooner.
In a three-year collaboration with a Japanese expert as well as the country's electronics giant NEC Corp., the Red Cross has developed a system involving drones and AI-aided equipment that can improve its own image analysis capability by learning about its surroundings.
The system will be particularly useful in countries such as Ethiopia and Ukraine, which have experienced heavy contamination from anti-personnel weapons, ICRC President Mirjana Spoljaric said.
Erik Tollefsen, head of the committee's weapon contamination unit, said, "In one day, the flying machine can do the work of a dog in six months."
"While a deminer will identify mines over an area of around 50 square meters in one day, the drone will be able to do it in four flights and will have done the same work over 100,000 sq meters," Tollefsen said.
Drones can take photos from several meters above the ground and even locate underground landmines through the use of infrared cameras.
Hideyuki Sawada, a professor of applied physics at Waseda University who is involved in the system's development, said using AI technology will help detect landmines more effectively and quickly, reducing mine-related deaths and injuries.
"Thanks to the development of the Deep Learning method, the recognition of images and the ability to detect and identify objects have drastically improved," Sawada said.
It is more difficult for drones to detect landmines in urban settings or jungles than in open fields or prairies.
But Sawada said that if AI learns more under such circumstances, its landmine detection capabilities will develop further.