Artificial intelligence learned to determine rocks by photos

Researchers from Skolkovo higher education institution science and technologies trained Neural networks effectively to distinguish samples of rocks in photos of boxes of a core. It allowed to accelerate process analyses to 20 times, and also to automate the core description. The developed algorithm is used in digital service of geological investigation DeepCore created in companies by Digital Petroleum. The press service told about work Skolkovo higher education institution science and technologies. One of routine problems of geological researches -- the description of rocks. Often samples are taken in the form of a core and develop in boxes.