Drones: The Future of Mine Detection

Landmines, with their precursors being deployed as early as the American Civil War in the 1800s, are still in use today. However, landmines do not become unarmed after a war, and will remain where they are, a constant threat, until proper removal by a professional team. Theoretically, only anti-tank landmines are permitted in war; however, when concerning certain countries, the Geneva Convention, which identifies the legal humanitarian treatment of people in war, is merely the Geneva suggestion (also consider that the US did not sign this specific treaty about anti-personnel landmines either). For instance, we still see the usage of anti-personnel landmines in the Russian-Ukraine conflict.

The reason that anti-personnel landmines were banned in the first place was due to the indiscriminate targeting of such a weapon. While anti-vehicle landmines only trigger when the weight of a vehicle is put on it, a civilian could accidentally activate an anti-personnel landmine. Of the more problematic ones, the PFM-1 is commonly employed by Soviet and Russian forces. One of the reasons that the PFM-1 specifically is also unethical is due to the fact that it is a scatter mine. This means that it is scattered through an area by either dropping a large amount of them from an aircraft (such as a helicopter or bomber) or even fired through a mortar. The PFM-1 mine is also frowned upon as it takes a very appealing appearance to children, often prompting them to pick the mines up. These mines can also explode if held between the thumb and index finger.

Obviously, clearing minefields after a war is a priority. However, it takes an incredible amount of time, especially going over an empty field. Hence why people came up with the idea of using drones to detect mines. These range from the more rudimentary designs such as the one Ukrainian teen, Igor Klymenko, created during the recent Russian invasion of Ukraine – which is essentially a drone with the important bits of a metal detector attached to it with a locator – to some more complex designs that involve using multiple scanners such as multispectrum and thermal scans.

The earliest of these designs was published in a journal article written in 2018 that involved only the usage of thermal scanners. This drone was coincidentally also designed to spot PFM-1 anti-personnel landmines (along with other explosives). It was found that the PFM-1 landmine would turn up on a thermal camera extremely clearly, with it being much colder than the surrounding environment.

The mines are relatively clear when scanned with a thermal camera (as seen in the red in the thermal picture). The mines and canisters have different specific heats as compared to the surrounding environment, causing them to change temperature differently. This means that when the temperature is increasing or decreasing, at sunrise and sunset, the mines and canisters would be at a different temperature than the surrounding (hence why it shows up on a thermal camera). The researchers also managed to achieve this using a 3DR Solo UAV (which is a typical commercial drone) with a FLIR Vue Pro R (A relatively cheap thermal camera), which adds up to a total of around $3000. To put this into perspective, an entry-level Ground penetrating radar costs upwards of $14,000 with more reliable ones costing around $25,000 to $35,000.

The drone that Igor Klymenko designed involved attaching a metal detector to a drone, before setting it on a specific flight path, with it recording detections along with the time detected. By inputting the time of detection into an algorithm that converts the time detected to coordinates relative to the starting position, Klymenko is able to detect where landmines are. Currently, there are two prototypes made with help from scientists and programmers and Klymenko received the Chegg.org Global award (as well as $100,000) for his innovation.

The latest design involved using multiple drones to scan over the same area, using different scanners. The basic idea is about the same as the first design with the different cameras’ data being used to find landmines, however, this time with a twist. The data is fed into a machine learning algorithm, specifically a convolutional neural network (which is great for image work), which then helps determine where mines are. (How the drone gets the coordinates is not specified in the source article). Currently, this is still in development, with more tests and improvements to the AI to be done.

These three drone systems would be a great help to demining efforts as it saves time and helps guarantee the safety of demining personnel. Currently, the rate of death and injury is one and two respectively out of 5000 mines cleared, which is still a significant number. Hopefully, these new drone systems can be soon implemented to lower the deaths and increase the efficiency of demining at lower costs.

Works cited:

Nikulin, A., De Smet, T. S., Baur, J., Frazer, W. D., & Abramowitz, J. C. (2018, October 23). Detection and identification of remnant PFM-1 ‘Butterfly mines’ with a UAV-based thermal-imaging protocol. MDPI. https://www.mdpi.com/2072-4292/10/11/1672#

Osborne, M. (2022, September 23). A Ukrainian teenager invents a drone that can detect land mines. Smithsonian Magazine. https://www.smithsonianmag.com/innovation/a-ukrainian-teenager-invents-a-drone-that-can-detect-land-mines-180980826/

Robinson, A., & Smith, D. (2022, September 7). To clear deadly land mines, science turns to drones and machine learning. Scientific American. https://www.scientificamerican.com/article/to-clear-deadly-land-mines-science-turns-to-drones-and-machine-learning/