1. School of Geophysics and Information Engineering, China University of Geosciences (Beijing), Beijing 100083, China 2. Command Center of Natural Resource Comprehensive Survey, China Geological Survey, Beijing 100083, China
Rapid progress has been made in the application of big data and artificial intelligence technology in the prediction of mineral resources. However, the application of machine learning technology based on convolutional neural networks remains in the exploration and experimental stages, with few practical examples and accomplishments achieved in the exploration and prediction of mineral resources in China. This study proposed applying convolutional neural networks to the exploration of gold deposits. Specifically, a neural network was trained for 2000 rounds using measured geological, mineral, geophysical, and geochemical data collected from a mineralization region covering an area of 3×104 km2 in a gold deposit in Shandong Province. Consequently, a 1D convolutional neural network model with accuracy of 0.95 and a loss rate of 0.11 was obtained. This model was employed to predict the distribution locations of gold deposits (exploration target areas) in other unknown areas in Shandong Province, yielding encouraging outcomes.
Yuan F, Li X H, Zhang M M, et al. A method for three-dimensional comprehensive information mineralization prediction of concealed ore bodies[J]. Journal of Geology, 2014, 88(4):630-643.
[3]
Harris D P. Undiscovered uranium resources and potential supply[C]// Workshop on Concepts of Uranium Resources and Producibility. Washington: Board Mineral Energy Resources,National Academy of Sciences, 1978:51-81.
[4]
Singer D A. Basic concepts in three-part quantitative assessments of undiscovered mineral resources[J]. Nonrenewable Resources, 1993, 2 (2):69-81.
doi: 10.1007/BF02272804
Zhao P D. Quantitative prediction and evaluation of "Three Linked" resources:A discussion on digital prospecting theory and practice[J]. Earth Science-Journal of China University of Geosciences, 2002, 27(5):482-489.
Zhang Y, Ma X D, Li Y J, et al. The application of deep learning to the shale gas content prediction in Nanchuan(South Sichuan)[J]. Geophysical and Geochemical Exploration, 2021, 45 (3):569-575.
Cheng Q M. Nonlinear ore-forming prediction theory:Multifractal singularity-generalized self similarity-fractal lineage model and method[J]. Earth Science-Journal of China University of Geosciences, 2006, 31(3):337-348.
Zhou Y Z, Li P X, Wang S G, et al. Research background and progress of big data and intelligent deposit models for mineral deposits[J]. Bulletin of Mineral and Rock Geochemistry, 2017, 36(2):327-331,344.
Liu Y P, Zhu L X, Zhou Y Z. Convolutional neural networks and their application in ore prospecting and prediction:A case study of the Zhaojikou lead-zinc deposit in Anhui Province[J]. Acta Petrologica Sinica, 2018, 34(11):3217-3224.