NEC has conducted a demonstration experiment to predict the area where landmines are buried using AI in order to reduce the number of casualties caused by landmines and contribute to safe, prompt, and efficient removal work. As a result, it was confirmed that the matching rate with the actual buried location was approximately 90%, and that the prediction could be performed with high accuracy.
This demonstration experiment is a co-creation project with the International Committee of the Red Cross (ICRC) that started in June 2009.
Landmines are difficult to locate in forests and deserts because protective colors are used. In addition, since it is buried in selected places where people and vehicles pass, it is easy to endanger human lives even after the conflict is over.
In 2021, 5,554 people were killed and injured worldwide by remnants of war, including cluster bombs and landmines, most of them civilians.
In addition, it is difficult to develop land in areas where landmines may be buried, hindering economic development.
In the demonstration experiment, topographical information such as rivers and mountainous areas, open data such as the location of factories and important buildings, and information from residents were provided by the ICRC, and NEC’s AI was able to select and analyze effective information. Predicted areas with a high probability of landmine burial.
Based on information about conflict areas in Asia, we have achieved a match rate of about 90% with the locations where landmines were actually buried.
Based on a huge amount of unorganized information, it is now possible to quickly identify areas with a high possibility of burying mines, which used to take many people a lot of time.