The Russian invasion of Ukraine has seen so many landmines deployed across the country that clearing them would take 700 years, say researchers. To make the task more manageable, Ukrainian scientists are turning to artificial intelligence to identify which regions are a priority for de-mining, though they expect some may simply have to be left as a permanent 鈥渟car鈥� on the country.
Russia鈥檚 minelaying and Ukrainian efforts to remove the explosives began with the initial invasion of Crimea in 2014, which saw a few hundred square kilometres of land contaminated. Now, two years after Russia鈥檚 wide-scale invasion, it is thought that up to 144,000 square kilometres could be mined. This is far more than could be cleared in many generations, potentially necessitating a semi-permanent exclusion zone such as the created by France after the first world war.
鈥淧eople get scars from war, and countries get scars from war,鈥� says , who is working as an advisor to the Ukrainian government as part of a team developing an AI model that is already assisting in de-mining.
Advertisement
The model considers vast amounts of data, including tax and property ownership records, agricultural maps, data on soil fertility, logs from the military and emergency services of where bombs and shells have landed, information gleaned from satellite images and interviews with local civilians and the military. Even climate change models and data on population density derived from mobile phone operators could be assessed. The AI then weighs factors such as civilian safety and potential economic benefits to determine the importance of a given piece of land and how urgent it is to make it safe.
, a deputy minister at Ukraine鈥檚 Ministry of Economy, is leading the team, and he likens the task of de-mining during an ongoing war to designing and building a submarine entirely underwater, except that the water is on fire. 鈥淚t鈥檚 a big problem,鈥� he says.
For example, says Bezkaravainyi, since agriculture is necessary for keeping Ukraine鈥檚 population fed as well as supporting the country鈥檚 finances 鈥� Ukraine is a major food exporter 鈥� it might make sense to prioritise agricultural land. But the problem isn鈥檛 so simple, he says.
Various mines and shells leave contamination that takes different amounts of time to clear, and craters can wipe away the fertile upper layers of soil and spread sand from beneath, ruining the ability to grow crops on the land in the short term. The AI uses satellite images to detect craters and take these contamination and fertility factors into account, not only finding agricultural land that needs restoration but also calculating how hard and lengthy those efforts will be, prioritising land that can be put back to crops the quickest.
快猫短视频 was shown photographs of one crater, taken somewhere in Ukraine, caused by a shell that hit a farm鈥檚 store of volatile ammonium nitrate fertiliser. Bezkaravainyi says that the 40-meter-wide, 20-metre-deep crater is now a 鈥渇uture lake鈥� rather than a field.
Once key areas are identified, the location is handed over to de-mining engineers who can conduct more intensive drone surveys at a rate of 3 hectares per hour. Quadcopters fitted with ground-penetrating radar and magnetometers are already finding some mines, although others are plastic and trickier to locate, and tall vegetation like corn or sunflowers can also affect accuracy.
Military demands on the de-mining engineers make the problem even more complex. While restoring land for future use, the same engineers must create mine-free corridors to allow troops to push forward, while further back from the front lines, engineers deliberately leave areas mined as a buffer. The AI suggestions will only ever be an approximation, says Pareniuk.
鈥淭here is no time to sit and try and solve all of the problems,鈥� she says. 鈥淲e have to construct something, plan it and troubleshoot it at the same time.鈥�
