zfbi/DeepISMNet: DeepISMNet: Three-Dimensional Implicit Structural Modeling with Convolutional Neural Network

Implicit structural modeling using sparse and unevenly distributed data is essential for various scientific and societal purposes ranging from natural source exploration to geological hazard forecasts. In this study, we propose an efficient deep learning method using a convolution neural network to predict a scalar field from sparse structural data associated with distinct stratigraphic layers and faults. This deep learning architecture is beneficial for the flexible incorporation of empirical geological knowledge when trained with numerous and realistic structural models that are automatically generated from a data simulation workflow. It also presents an impressive characteristic of integrating various types of structural constraints by minimizing a hybrid loss function between the models being compared, opening new opportunities for further improving geological structural modeling. Moreover, the deep neural network, after training, is highly efficient to predict implicit structural models in real-world applications.

As described in Three-Dimensional Implicit Structural Modeling Using Convolutional Neural Network by Zhengfa Bi¹, Xinming Wu¹, Zhaoliang Li², Dekuan Change³ and Xueshan Yong³. ¹University of Science and Technology of China; ²China Aero Geophysical Survey and Remote Sensing Center for Natural Resources; ³Research Institute of Petroleum Exploration & Development-NorthWest(NWGI), PetroChina.