Improving Vertical Resolution of Vintage Seismic Data by a Weakly Supervised Method Based on Cycle Generative Adversarial Network

Dawei Liu,Wenli Niu,Xiaokai Wang,Mauricio D. Sacchi,Wenchao Chen,Cheng Wang

DOI: https://doi.org/10.1190/geo2023-0006.1

IF: 3.264

2023-01-01

Geophysics

Abstract:Seismic vertical resolution is critical for accurately identifying subsurface structures and reservoir properties. Improving the vertical resolution of vintage seismic data with strongly supervised deep learning is challenging due to scarce or costly labels. To remedy the label-lacking problem, we develop a weakly supervised deep-learning method to improve vintage seismic data with poor resolution by extrapolating from nearby high-resolution seismic data. Our method uses a cycle generative adversarial network with an improved identity loss function. In addition, we contribute a pseudo-3D training data construction strategy that reduces discontinuity artifacts caused by accessing 3D field data with a 2D network. We determine the feasibility of our method on 2D synthetic data and achieve results comparable to the classic time-varying spectrum whitening method on field poststack migration data while effectively recovering more high-frequency information.