A Deep Learning-based time shift objective function for Full Waveform Inversion

TL;DR

This paper introduces a neural network-based time shift objective function for Full Waveform Inversion, improving stability and efficiency over traditional methods like DTW, and demonstrating comparable convergence on synthetic datasets.

Contribution

The paper proposes a neural network approach to learn time shifts for FWI, offering a stable, differentiable, and computationally efficient alternative to DTW-based misfit functions.

Findings

Achieves similar convergence to SoftDWT in synthetic tests

Reduces computational time of adjoint source calculation

Demonstrates effectiveness on Marmousi and Chevron datasets

Abstract

Full Waveform Inversion (FWI) is a powerful technique for estimating high-resolution subsurface velocity models by minimizing the discrepancy between modeled and observed seismic data. However, the oscillatory nature of seismic waveforms makes point-wise discrepancy measures highly prone to cycle-skipping, especially when the initial velocity model is inadequate. To address this challenge, various alternative misfit functions have been proposed in the literature, each with unique strengths and limitations. Dynamic Time Warping (DTW) is a popular technique in signal processing for aligning time series using dynamic programming. While a differentiable variant of DTW has been recently proposed, its use in FWI is hindered by high-frequency artifacts in the adjoint source and the substantial computational cost of gradient evaluations. In this study, we propose a neural network-based approach to learn the time shifts that align two time series in a supervised manner. The trained network is then utilized to compare traces from observed and modeled seismic data, offering a stable and computationally efficient alternative to DTW. Furthermore, the inherent differentiability of neural networks via backpropagation enables seamless integration into the FWI framework as a misfit function. We validate this approach on two synthetic datasets, namely the Marmousi model and the Chevron blind test dataset, demonstrating in both cases a similar convergence behavior to that of SoftDWT whilst drastically reducing the computational time of the adjoint source calculation.