Innopolis Develops Neural Network for Geological Core Analysis
The system’s performance is already approaching the level of an experienced specialist.
Innopolis University has patented a new AI system designed for mining geomechanics. The platform scans photographs of geological core samples and automatically identifies geological structures within them. Experts say the technology could significantly improve the accuracy of mineral exploration while speeding up the creation of geological models for mining deposits.
“Existing software platforms are either too specialized or require constant oversight from experts. Our method addresses that problem using artificial intelligence,” said Arseny Pinigin, head of AI technologies at the Center for Industrial and Engineering Solutions at Innopolis University.
The system currently operates with an accuracy rate of about 70%, a level developers say is comparable to that of an experienced geologist. The technology is expected to continue improving as development progresses.
