Data-Driven Seismic Waveform Inversion: A Study on the Robustness and Generalization

Abstract

Acoustic-and elastic-waveform inversion is an important and widely used method to reconstruct subsurface velocity image. Waveform inversion is a typical non-linear and ill-posed inverse problem. Existing physics-driven computational methods for solving waveform inversion suffer from the cycle skipping and local minima issues, and not to mention solving waveform inversion is computationally expensive. In recent years, data-driven methods become a promising way to solve the waveform inversion problem. However, most deep learning frameworks suffer from generalization and over-fitting issue. In this paper, we developed a real-time data-driven technique and we call it VelocityGAN, to accurately reconstruct subsurface velocities. Our VelocityGAN is built on a generative adversarial network (GAN) and trained end-to-end to learn a mapping function from the raw seismic waveform data to the velocity image. Different from other encoder-decoder based data-driven seismic waveform inversion approaches, our VelocityGAN learns regularization from data and further impose the regularization to the generator so that inversion accuracy is improved. We further develop a transfer learning strategy based on VelocityGAN to alleviate the generalization issue. A series of experiments are conducted on the synthetic seismic reflection data to evaluate the effectiveness, efficiency, and generalization of VelocityGAN. We not only compare it with existing physics-driven approaches and data-driven frameworks but also conduct several transfer learning experiments. The experiment results show that VelocityGAN achieves state-of-the-art performance among the baselines and can improve the generalization results to some extent.

Neural Networks for Seismic Data Inversion

“…In the intention of reducing the training time of the GAN, the generator network was preconditioned with a training for model generation without considering the discriminator. With this transfer learning strategy, based on [3], the initial outputs of the generator will not be completely random. As the goal of preconditioning is only to get a better initial state, the training was done with a few epochs, generating the results in Figure 4.…”

Wavefield Reconstruction Inversion via Physics-Informed Neural Networks

“…Convolutional NN (CNN) has strong capabilities in extracting features from a large number of images, and it has been effectively applied to detect salt bodies [32], horizons, and faults from seismic images [33], predict low-frequency components from highfrequency data [34]. Besides applications for analyzing the image features, CNN can also achieve seismic wave simulation in complex media [35] and direct velocity inversions [36], [37], [38]. [39] implement FWI using a physics-based NN, but this method still suffers from the cycle-skipping issue in the conventional FWI.…”

SeismoGen: Seismic Waveform Synthesis Using GAN With Application to Seismic Data Augmentation

“…Neither of these studies explores data augmentation techniques for seismic event detection, as their focus primarily relates to feature extraction using GAN. Generative models have also been proved to be effective in other geophysical applications such as inversion (Z. Zhang & Lin, 2020; Zhong et al., 2020), data processing (Picetti et al., 2019), interpretation (Lu et al., 2018), and many others.…”