17.01.2025.

Sensors, drones, and artificial intelligence: How innovations accelerate land demining in Ukraine

Ukraine is currently the most mined country in the world. According to the Ministry of Economy of Ukraine, as of the end of 2024, a total of 138,500 square kilometres of land was potentially contaminated. With traditional demining methods, it would take almost a century to clear the land of explosive ordnance.

However, innovative remote explosive detection technologies can significantly speed up this process. Ukrainian developers are already creating unique solutions that combine different types of sensors with artificial intelligence to search for dangerous objects effectively.

In the first article of the special partnership project between LIGA.net and the United Nations Development Programme (UNDP) in Ukraine, we described how the new humanitarian demining market was launched in Ukraine and how international partners are helping the country in this field.

Our second article focuses on the development of demining technologies and how they can accelerate the demining of Ukrainian territories.

Sensor technologies in humanitarian demining

In May 2024, UNDP conducted a large-scale test of various types of sensors for detecting explosive devices. The test results showed that high-resolution cameras, thermal sensors, and magnetometer-based technologies are the most effective.

Thermal sensors
These are devices that "see" heat. They can detect objects based on temperature, even if they are not visible to the naked eye. For example, a landmine or other explosive object may heat up under the sun and then emit heat that a thermal sensor can recognize. This technology works well for detecting objects on the surface, but its effectiveness may decrease if the mine is heavily concealed by vegetation or soil.

Each type of sensor has its advantages and limitations. High-sensitivity optical cameras mounted on drones are effective at detecting objects on the surface. However, their use can be hampered by vegetation or other obstacles. Magnetometers are capable of detecting mines and other unexploded ordnance containing metal, but their effectiveness is reduced by the presence of debris, including metal, in the soil and the high mineralization of Ukrainian land. Detecting plastic mines with minimal metal content is also particularly challenging.

Magnetometers
These are devices that detect metal objects by measuring magnetic fields. The metal in mines or explosives affects this field, and the magnetometer "senses" these changes. They are useful for locating metallic explosive objects both on the surface and below ground. However, they can confuse dangerous objects with metallic debris in the soil, so they are not always accurate.

"We’re now coming to the point where we need to develop technologies that combine different sensors, as there’s no universal tool that can work with any type of explosive object," explains UNDP mine action specialist Edward Crowther.

Remote detection technologies for hazardous objects are being actively developed in Ukraine. In particular, the State Emergency Service (SES) is using satellite imagery to identify potentially contaminated areas, which allows for more efficient demining planning and prioritization. Most humanitarian demining operators use drones with optical cameras and artificial intelligence systems for image processing. Artificial intelligence can significantly speed up the data analysis process.

"So far, artificial intelligence has been used mainly to process visual data," says Crowther. "For example, when a drone flies over a certain area that is suspected of being contaminated, it takes hundreds, perhaps thousands of different pictures. Then they need to be checked for signs of certain explosive objects – anti-personnel mines, for example. (The use of AI) leads to a significant increase in productivity in processing such images, because an AI-based machine can process these images much faster and more accurately than a human."

At the same time, Crowther emphasizes that despite all the successes of artificial intelligence, a human being must always be included in the data processing process to verify the information. After all, humanitarian demining is primarily about ensuring human safety.

"In August, UNDP, together with the Ministry of Economy, trained ten veterans with disabilities to become visual data analysts," Crowther adds. "Seven of them are already working in their speciality, and five of them are employed by The HALO Trust (the largest international non-profit organization engaged in humanitarian demining – ed.), where they work on data analysis. This not only helps with demining but also facilitates the reintegration of veterans into civilian life by giving them the opportunity to apply their skills to an important cause."

According to Crowther, Ukraine has become a testing ground for the development of innovative demining technologies. By combining international experience and local innovations, effective solutions are being created to overcome unprecedented challenges in the field of demining.

"The developments of Ukrainian specialists and methodologies created to address humanitarian demining in Ukraine will change this area at the global scale," Crowther says. "The technologies being developed in Ukraine will undoubtedly be used in this activity around the world and will lead to significant changes."

Testing sensor technologies in Ukraine

In Ukraine, experts from various fields, including mining, IT, and even archaeology, are applying their knowledge to the development of technologies for remotely detecting explosive ordnance. Their proposals, and the ideas of international companies, are being tested on training grounds and under real conditions in the field. Even at this early stage, some of the results are already impressive.

In particular, during the aforementioned tests in May 2024, in which seven Ukrainian and international teams participated, some companies managed to detect 78% of the hazardous items on the surface, about 70% of those underground, and 62% of the plastic items.

According to Kateryna Drozd, project manager for innovation at the Mine Action Support Team at the Ministry of Economy of Ukraine, this summer tests were also conducted in real conditions in the Kamyanka community in Kharkiv Oblast, which was under occupation for about five months in 2022. The territory was heavily contaminated with PFM-1 anti-personnel mines, which are difficult to detect due to their plastic body and small size.

"We do a lot of testing," Drozd notes. "The best companies that show good results are then invited to carry out real work and collaborate with various demining operators."

At the same time, for such technologies to work quickly and correctly, the artificial intelligence that processes the images must be able to recognize explosive objects quickly and correctly. The Ministry of Economy is currently working on a database of common explosive items that will help developers train their technologies.

"You can combine sensors, but if there’s no database, it won’t work, unfortunately," explains Drozd. "We can't just take photos from the Internet and train an artificial intelligence, because there are a lot of modifications, and more of them appear every day. That's why lots of Ukrainian developers are working with the State Emergency Service, the State Special Transport Service, and operators in the fields so they can see everything live and create this database."

At the same time, standardization and certification are serious challenges for Ukraine in developing sensor technologies for detecting dangerous objects. As of today, there are no international standards for the use of sensor technologies to detect explosive devices – evidence is still being collected to determine in which cases certain technologies can be used.

"We have to remember that humanitarian demining is, first and foremost, about people's lives, and we cannot certify equipment that provides only 78