.png){kind=icon}
Return
Intelligent Pre-Stack Common Offset Gather Denoising With Post-stack Structure Guidance
.png){kind=icon}
PRE
AI
DOI：10.1109/tgrs.2026.3738458.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Seismic imaging is primarily used for subsurface exploration. However, the raw imaging results are often severely contaminated by various noises. To eliminate the influence of noises, various seismic denoising algorithms have been developed. Recently, deep-learning-based seismic denoising has been extensively investigated and has demonstrated significant potential. Nevertheless, most of them take only the noisy seismic image as input and directly produce the denoising output. For Pre-stack imaging gathers with extremely strong noise interference, it becomes highly challenging to discern the true structure of effective signals from the noisy input alone. This can lead denoising networks to misinterpret high-amplitude structured noise as genuine signal components, resulting in erroneous restoration. This study focuses on the Pre-stack common offset gather (COG) denoising task and proposes the prior-guided denoising (PGD) algorithm to address this challenge. The proposed PGD algorithm leverages the corresponding stacked image as a structural prior, automatically extracting guidance information to inform the denoising inference for each offset. The advantages of the proposed PGD algorithm over existing algorithms are validated through both synthetic experiments and field data applications. Field application results demonstrate that, by utilizing the stacked image as structural guidance, the PGD algorithm can restore clean single-offset images comparable in quality to Post-stack images. This not only delivers high-fidelity data for subsequent amplitude variation with offset (AVO) analysis but also enables fine-scale interpretation based on Pre-stack imaging gathers.
Keywords:
Common offset gather (COG)
deep-learning-based denoising
prior-guided inference
seismic denoising
Journal
IF:
8.6
Papers:
2.2W
Citations:
10.7W
Organization
Cited Papers
No cited papers available