3DInception-U: Lightweight Network for 3-D Magnetotelluric Inversion Based on Inception Module

Abstract

In the field of geophysical exploration, the application of deep learning techniques has garnered significant attention. This letter proposes a new deep learning model for 3-D magnetotelluric inversion, named 3DInception-U. In this model, we integrate the inception module into the network architecture and combine the concatenation layer with a U-Net structure. This model has two advantages: First, the inception module, along with the deep concatenation layer, enhances the network’s capability for feature extraction and representation, and second, the skip connections in the U-Net facilitate information propagation, enabling the design of a network with fewer parameters but better performance. We produced 10 000 3-D complex samples for training by Gaussian random fields (GRFs) and compared 3DInception-U with existing 3-D magnetotelluric (MT) inversion models and applied it to real geological interpretation. The results demonstrate that this network architecture achieves good inversion accuracy and robustness.