Time-Domain Elastic Full Waveform Inversion With Frequency Normalization

Fang, Jinwei; Zhou, Hui; Li, Yunyue Elita; Shi, Ying

doi:10.1109/tgrs.2023.3246772

Cited by 2 publications

(3 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The important component proposed by Wang et al. (2023) is to replace the denoising process in PnP‐FISTA with an intermediate denoising sampling of a diffusion model. The connection between these approaches is highlighted by comparing Algorithms 2 (Table 1) and 4 (Table 2).…”

Section: Theorymentioning

confidence: 99%

“…Table 2 further compares these two workflows at a high level. Unlike the mono‐parameter acoustic version (Wang et al., 2023), solving the elastic equation requires dealing with three model parameters ( V p , V s , ρ ). Storing the relation between these parameters as prior, especially their structural consistency, is often crucial.…”

Section: Theorymentioning

confidence: 99%

“…(2017) later extended this idea to work on nonlinear imaging problems. In this work, we promote using a recent diffusion model‐based adaptive regularization framework (Wang et al., 2023) to solve the multi‐parameter nonlinear inversion. Our regularization framework naturally accounts for the coupled relationship of the elastic parameters while constraining the individual elastic parameters.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learned Regularizations for Multi‐Parameter Elastic Full Waveform Inversion Using Diffusion Models

Taufik,

Wang,

Alkhalifah

2024

Journal of Geophysical Research: Machine Learning and Computati

View full text Add to dashboard Cite

Elastic full waveform inversion (EFWI) promises to account for the Earth's elastic nature and corresponding reflectivity, which is often disregarded in the commonly used acoustic FWI. However, EFWI usually requires a more sophisticated recording apparatus (beyond the usual single‐component data). Even in the presence of multicomponent recordings, an empirical formulation that relates the elastic parameters is usually employed. Such approximations, thus, render the inverted elastic parameters (and their relationship) hostage to our assumptions. To overcome these limitations, we introduce learned regularization using diffusion models. Specifically, we first train the (unsupervised) diffusion model to understand the coupling relationship of the distribution of the elastic parameters and use the trained model in the inversion process with a negligible additional computational cost. To fully realize the effect of our regularization and to mimic a realistic scenario, the vertical component of the particle velocity is used to invert the elastic parameters. Unlike other learned (deep) regularizers, diffusion models offer a unique conditional capability that suits the nature of an FWI process (going from low to high frequency) while maintaining the problem‐agnostic feature of such regularizers. Numerical experiments, ranging from synthetic to land field data, show that our framework solves the illumination effects from an imperfect acquisition setup and provides more realistic elastic parameter ratios than the conventional EFWI. We also empirically demonstrate that, unlike traditional regularization schemes, our framework converges to better model estimates that fit the observed data better.

show abstract

Section: Theorymentioning

confidence: 99%

Section: Theorymentioning

confidence: 99%