In this paper, we developed an innovative machine learning (ML) method to determine salt structures directly from gravity data. Based on a U-net deep neural network, the method maps the gravity downward continuation volume directly to a salt body mask volume, which is easily interpretable for an exploration geophyisicist. We also studied the feasibility to apply the method to different gravity field data. We conclude that the ML based method from gravity data complements seismic data processing and interpretation for subsurface exploration. For the region where no or limited seismic data are available, this ML based salt identification can save iterative efforts (>50%) in the conventional gravity inversion process and identify major salt bodies in the region.

Keywords:
neural network, machine learning, deep learning, artificial intelligence, deep neural network, prediction, new exploration area, application, seismic survey line, dice coefficient
Subjects:
Reservoir Characterization, Information Management and Systems, Seismic processing and interpretation, Artificial intelligence, Neural networks
Copyright 2020, Society of Petroleum Engineers
You can access this article if you purchase or spend a download.