Learning 3D mineral prospectivity from 3D geological models using convolutional neural networks: Application to a structure-controlled hydrothermal gold deposit

Three-dimensional (3D) geological models are typical data sources in 3D mineral prospectivity modeling. However, identifying prospectivity-informative predictor variables from is a challenging and work-intensive task. Motivated by the ability of convolutional neural networks (CNNs) to learn intrinsic features, this paper, we present novel method that leverages CNNs models. By exploiting learning ability, proposed simplifies complex correlations mineralization circumvent need for designing variables. Specifically, analyze unstructured using CNNs—whose inputs should be structured—we develop 2D CNN framework where geometry boundary compiled reorganized into multi-channel images fed CNN. This ensures effective efficient training while facilitating representation control. The presented applied structure-controlled hydrothermal deposit, Dayingezhuang gold deposit eastern China; compared with modeling methods designed results show has performance boost terms decreases workload prospecting risk prediction deep-seated orebodies.