GMLAN: Grouped-residual and multi-scale large-kernel attention network for seismic image super-resolution

Abstract

The resolution of seismic images significantly impacts the accuracy of subsequent seismic interpretation and reservoir location. However, the resolution of seismic images often degrades due to the influence of multiple factors, making super-resolution of seismic images essential and critical. We propose a grouped-residual and multi-scale large-kernel attention network (GMLAN) framework, trained on synthetic seismic images to achieve excellent seismic image super-resolution on field seismic data. GMLAN is primarily composed of two modules: The feature extraction module (FEM) and the image reconstruction module (IRM). The FEM consists of two components: Shallow feature extraction (SFE) and deep feature extraction (DFE). The SFE component is designed to capture the basic information of seismic images, such as large-scale structures and morphological features of the strata. The DFE component serves as the cornerstone of the feature extraction process, leveraging residual groups and multi-scale large-kernel attention to distill detailed features from seismic images, such as stratigraphic interfaces, dip angles, and relative amplitudes. Finally, the IRM utilizes sub-pixel convolution, a learnable upsampling technique, to reconstruct super-resolution seismic images while preserving the continuity of seismic features. The framework demonstrates satisfactory performance on both synthetic and field data.

Keywords

Seismic images / Super-resolution / Deep learning / Grouped-residual structures / Malti-scale large-kernel self-attention

Funding

This work was supported by BGP’s Science and Technology project (01-04-02-2024).

References

[1] Soubaras R, Dowle R, Sablon R. Broadseis: Enhancing interpretation and inversion with broadband marine seismic. CSEG Recorder. 2012: 37(7):40-46.

[2] Rebert T, Sablon R, Vidal N, Charrier P, Soubaras R. Improving pre-salt imaging with variable-depth streamer data. In: SEG Technical Program Expanded Abstracts 2012. United States: Society of Exploration Geophysicists; 2012. p. 1-5. doi:10.1190/segam2012-1067.1

[3] Wang Y, Wang J, Wang X, Sun W, Zhang J. Broadband processing key technology research and application on slant streamer. International Geophysical Conference, Beijing, China, 24-27 April 2018. Society of Exploration Geophysicists and Chinese Petroleum Society. United States: Society of Exploration Geophysicists; 2018. p. 135-138. doi:10.1190/IGC2018-034

[4] Zhang YG, Wang Y, Yin JJ. Single point high density seismic data processing analysis and initial evaluation. Shiyou Diqiu Wuli Kantan Oil Geophys. Prospect. 2010; 45:201-207. doi:10.13810/j.cnki.issn.1000-7210.2010.02.008

[5] Xiao F, Yang J, Liang B, et al. High-density 3D point receiver seismic acquisition and processing - a case study from the Sichuan Basin, China. First Break. 2014; 32(1):81-90. doi:10.3997/1365-2397.32.1.72598

[6] Zhang H, Bao X, Zhao H, et al. High-precision deblending of 3-D simultaneous source data based on prior information constraint. IEEE Geosci Remote Sens Lett. 2025; 22:1-5. doi:10.1109/LGRS.2025.3526972

[7] Shang XM, Diao R, Feng YP, Zhao CX. The application of spectral modeling method to high resolution processing of seismic data. Geophys Geochem Explor. 2014; 38(1):75-80. doi:10.11720/j.issn.1000-8918.2014.1.13

[8] Wang D, Yuan S, Liu T, Li S, Wang S. Inversion-based non-stationary normal moveout correction along with prestack high-resolution processing. J Appl Geophy. 2021; 191:104379. doi:10.1016/j.jappgeo.2021.104379

[9] Wu X, Ma J, Si X, et al. Sensing prior constraints in deep neural networks for solving exploration geophysical problems. Proc Natl Acad Sci U S A. 2023; 120(23):e2219573120. doi:10.1073/pnas.2219573120

[10] Mousavi SM, Beroza GC, Mukerji T, Rasht- Behesht M. Applications of deep neural networks in exploration seismology: A technical survey. Geophysics. 2023; 89(1):WA95-WA115. doi:10.1190/geo2023-0063.1

[11] Yuan S, Yu Y, Sang W, Xie R, Zhou C, Chen S. Seismic horizon picking using deep learning with multiple attributes. IEEE Trans Geosci Remote Sens. 2025; 63:1-16. doi:10.1109/TGRS.2025.3581462

[12] Zeng D, Xu Q, Pan S, Song G, Min F. Seismic image super-resolution reconstruction through deep feature mining network. Appl Intell. 2023; 53(19):21875-21890. doi:10.1007/s10489-023-04660-y

[13] Zhou R, Zhou C, Wang Y, Yao X, Hu G, Yu F. Deep learning with fault prior for 3-D seismic data super-resolution. IEEE Trans Geosci Remote Sens. 2023; 61:1-16. doi:10.1109/TGRS.2023.3262884

[14] Li J, Wu X, Hu Z. Deep learning for simultaneous seismic image super-resolution and denoising. IEEE Trans Geosci Remote Sens. 2022; 60:1-11. doi:10.1109/TGRS.2021.3057857

[15] Min F, Wang L, Pan S, Song G. D2UNet: Dual decoder U-net for seismic image super-resolution reconstruction. IEEE Trans Geosci Remote Sens. 2023; 61:1-13. doi:10.1109/TGRS.2023.3264459

[16] Liang J, Cao J, Sun G, Zhang K, Van Gool L, Timofte R. SwinIR: Image Restoration using Swin Transformer. 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). Montreal, BC, Canada. 2021. p. 1833-1844. doi:10.1109/ICCVW54120.2021.00210

[17] Zhong T, Zheng K, Dong S, Tong X, Dong X. Enhancing the resolution of seismic images with a network combining CNN and transformer. IEEE Geosci Remote Sens Lett. 2025; 22:1-5. doi:10.1109/LGRS.2024.3495659

[18] Zhong T, Yang F, Dong X, Dong S, Luo Y. SHBGAN: Hybrid bilateral attention GAN for seismic image super-resolution reconstruction. IEEE Trans Geosci Remote Sens. 2024; 62:1-12. doi:10.1109/TGRS.2024.3492142

[19] Lin L, Zhong Z, Cai C, Li C, Zhang H. SeisGAN: Improving seismic image resolution and reducing random noise using a generative adversarial network. Math Geosci. 2024; 56(4):723-749. doi:10.1007/s11004-023-10103-8

[20] Xiao Y, Li K, Dou Y, Li W, Yang Z, Zhu X. Diffusion models for multidimensional seismic noise attenuation and superresolution. Geophysics. 2024; 89(5):V479-V492. doi:10.1190/geo2023-0676.1

[21] Yuan S, Xu Y, Xie R, Chen S, Yuan J. Multi-scale intelligent fusion and dynamic validation for high-resolution seismic data processing in drilling. Pet Explor Dev. 2025; 52(3):680-691. doi:10.1016/S1876-3804(25)60596-9

[22] Zhou G, Zhi H, Gao E, et al. DeepU-Net: A parallel dual-branch model for deeply fusing multiscale features for road extraction from high-resolution remote sensing images. IEEE J Sel Top Appl Earth Obs Remote Sens. 2025; 18:9448-9463.doi:10.1109/JSTARS.2025.3555636

[23] Zhou Y, Li Z, Guo CL, Bai S, Cheng MM, Hou Q. SRFormer: Permuted self-attention for single image super-resolution. In: 2023 IEEE/CVF International Conference on Computer Vision (ICCV). United States: IEEE; 2023. p. 12734-12745. doi:10.1109/ICCV51070.2023.01174

[24] Li Y, Deng Z, Cao Y, Liu L. GRFormer:Grouped residual self-attention for lightweight single image super-resolution. In: Presented at: Proceedings of the 32^ndACM International Conference on Multimedia; 2024; Melbourne VIC, Australia. United States: Cornell University. doi:10.1145/3664647.3681554

[25] Yan Z, Zi-Xin W, Lin-qI C, Hong-Li D. Research on microseismic event localization based on convolutional neural network. JSE. 2024; 33(6):1-32.

[26] Xu Y, Yuan S, Zeng H, et al. Frequency-dependent multiscale network for seismic high-resolution processing. Geophysics. 2025; 90(4):V297-V312. doi:10.1190/geo2023-0682.1

[27] Wang Y, Li Y, Wang G, Liu X. Multi-Scale Attention Network for single Image Super-Resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; 2024. p. 5950-5960.

[28] Wu X, Geng Z, Shi Y, Pham N, Fomel S, Caumon G. Building realistic structure models to train convolutional neural networks for seismic structural interpretation. Geophysics. 2019; 85(4):WA27-WA39. doi:10.1190/geo2019-0375.1

[29] Zhao H, Gallo O, Frosio I, Kautz J. Loss functions for image restoration with neural networks. IEEE Trans Comput Imaging. 2017; 3(1):47-57. doi:10.1109/TCI.2016.2644865

[30] Lou Y, Zhang B, Wang R, Lin T, Cao D. Seismic fault attribute estimation using a local fault model. Geophysics. 2019; 84(4):O73-O80. doi:10.1190/geo2018-0678.1

[31] Yan B, Wang T, Ji Y, Yuan S. Multidirectional coherence attribute for discontinuity characterization in seismic images. IEEE Geosci Remote Sens Lett. 2022; 19:1-5. doi:10.1109/LGRS.2022.3151686