1887

Abstract

Summary

Seismic data is crucial for subsurface structural interpretation but is often compromised by noise. This paper presents a deep diffusion probabilistic model (DDPM) specifically designed for seismic denoising, complemented by an innovative signal-fitting training strategy that addresses the unique characteristics of seismic data. Unlike traditional denoising methods, which rely on rigid physical assumptions, our approach utilizes deep learning to adapt effectively to the complex features inherent in seismic data. We demonstrate that the DDPM, combined with the signal-fitting training strategy, outperforms conventional techniques in seismic image denoising tasks. Furthermore, we enhance the model’s generalization capability by incorporating regularization constraints into the loss function. Experimental results on the dataset indicate a significant improvement in signal-to-noise ratio and data fidelity, highlighting the superior performance of the proposed method.

© EAGE Publications BV
Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.202639117
2026-03-09
2026-02-19

  • From This Site

    /content/papers/10.3997/2214-4609.202639117
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Journal -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

References

  1. GulunayN.FXDECON and complex wiener prediction filter//SEG Technical Program Expanded Abstracts1986. Society of Exploration Geophysicists, 1986:279–281. https://doi.org/10.1190/1.1893128
     [Google Scholar]
  2. HoJ., JainA., AbbeelP.2020. Denoising diffusion probabilistic models, 34th Proc. Adv. Neural Inf. Process. Syst. (NIPS). Curran Associates Inc.; Vancouver, BC, Canada: 6840–6851
     [Google Scholar]
  3. HuangW., WangR., ChenY. et al., (2016) Damped multichannel singular spectrum analysis for 3D random noise attenuation. Geophysics, 81: V261–V270. https://doi.org/10.1190/geo2015-0264.1
     [Google Scholar]
  4. LiH., ShiJ., LiL. et al., (2022) Novel Wavelet Threshold Denoising Method to Highlight the First Break of Noisy Microseismic Recordings. IEEE Transactions on Geoscience and Remote Sensing, 60: 1–10. https://doi.org/10.1109/TGRS.2022.3142089
     [Google Scholar]
  5. WangF., ChenS. (2019) Residual Learning of Deep Convolutional Neural Network for Seismic Random Noise Attenuation. IEEE Geoscience and Remote Sensing Letters, 16: 1314–1318. https://doi.org/10.1109/LGRS.2019.2895702
     [Google Scholar]
  6. ZhongT., ChengM., DongX. et al., (2022) Seismic random noise suppression by using deep residual U-Net. Journal of Petroleum Science and Engineering, 209: 109901. https://doi.org/10.1016/j.petrol.2021.109901
     [Google Scholar]
/content/papers/10.3997/2214-4609.202639117
Loading
A New Training Strategy for the Diffusion Model in Seismic Denoising
Conference Proceedings 2026, 1 (2026); https://doi.org/10.3997/2214-4609.202639117
/content/papers/10.3997/2214-4609.202639117
/content/papers/10.3997/2214-4609.202639117
Loading

Data & Media loading...

Most Cited This Month Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error