oa Assisted processing of geological data with Deep Learning technologies: levers to optimize mineral exploration workflows?

The use of Deep Learning technologies to analyze unstructured data, for instance images and texts, has recently known significant improvement and democratization. As a contribution to the digital transformation of the mining industry, we present three practical use cases where these technologies were successfully applied to geological data. They come from a petroleum exploration context, but illustrate the potential of Deep Learning to optimize geological activities. In this study we aim at discussing how much these business cases can relate to mineral exploration, and whether the Deep Learning revolution could also benefit mineral exploration workflows. The first case is a lithological classification of macroscopic rock samples pictures, which could be extended to automated core interpretation and foreshadows autonomous mining robots. The second case is a delineation of micro-fossils on thin sections scans, which could be adjusted for accelerated detection of mineral grains or metallic nuggets. The third case is a scientific texts mining based on an entity extraction engine, which could be used for assisted deposit characterization from geological literature. Eventually, we conclude that fantastic innovation opportunities lie in the integration of Deep Learning technologies into mineral exploration workflows.