Deep Learning-based Seismic Reflectivity Estimation by Pre-training on Labeled Synthetic Data and Physics-guided Fine-tuning in Field Data

Clicks: 24
ID: 324344
2026
Free PDF
Article Quality & Performance Metrics
Overall Quality
Not rated
Combines reader engagement with the AI quality analysis. This article has not been analysed, so there is no overall score — reader engagement is measured and shown alongside.
AI Quality Assessment
Not analyzed
Readership in this journal
Steady

Ranked #140 of 345 articles by views in geophysical journal international

Most read Least read

Bar heights use a square-root scale. Only the 120 most-read articles are drawn; the journal has 345 in total.

Mint this article as an NFT
Not yet minted

Create a permanent, verifiable on-chain record of this article on the Scimatic Network. The NFT is held in your Journament account, and you can withdraw it to your own wallet at any time.

5 SUSD one-off · no wallet required
Sign in to mint
Abstract
Summary Reflectivity estimation aims to enhance the resolution of seismic data, providing crucial support for the detailed inversion of reservoir parameters. We propose a method that combines supervised pre-training with synthetic data and physics-guided fine-tuning in field data to estimate reasonable reflectivity from seismic data. Initially, a U-shaped network is pre-trained by supervised learning on a large amount of synthetic seismic data. Subsequently, multiple geophysically meaningful constraints including structure-oriented smoothness, reflectivity sparsity, and data reconstruction, are introduced to formulate a self-supervised or unsupervised learning mechanism to fine-tune the pre-trained model so that it is better adapted to field data for obtaining more reasonable reflectivity. The pre-trained model provides an initial reflectivity that aligns with fundamental structural features. Based on the initial estimate, the model is further optimized through physics-guided fine-tuning to obtain a more reasonable reflectivity estimation that better reflects the characteristics of the field data and geophysical priors. Furthermore, the well-log-based correlation evaluation metric is applied to automatically and adaptively determine the early-stopping point of the fine-tuning process where favorable results are achieved. Experiments on synthetic and field seismic data confirm that the proposed method yields reasonable and high-resolution reflectivity estimation.
Reference Key
openalex_W7201960120 Use this key to autocite in the manuscript while using SciMatic Manuscript Manager or Thesis Manager
Authors Yuting Wang, Jintao Li, Xiaoming Sun, Xinming Wu
Are you one of these authors? Claim your profile
Journal geophysical journal international
Year 2026
DOI
10.1093/gji/ggag317
URL
https://doi.org/10.1093/gji/ggag317
Keywords Keywords not found

Citations

No citations found. To add a citation, contact the admin at info@scimatic.org

Comments

No comments yet. Be the first to comment on this article.