Multi-stage progressive network for seismic random noise suppression

Abstract

Seismic data quality frequently deteriorates due to random noise contamination, substantially impeding subsequent processing and geological interpretation. While deep learning approaches have emerged as powerful tools for noise suppression, conventional single-stage architectures exhibit inherent limitations in handling complex seismic features while preserving subtle geological details. These challenges motivate the development of advanced multi-stage neural networks for seismic data enhancement. The proposed multi-stage progressive U-shaped convolutional network (MPU-Net) architecture addresses these limitations through supervised cross-stage attention mechanisms that maintain feature connectivity throughout the network. Building upon this foundation, group enhanced convolutional blocks (GEB)-MPU-Net introduces GEB to specifically counteract the progressive attenuation of shallow features in deep networks. This dual-stage enhancement strategy combines hierarchical feature preservation, adaptive information fusion, and stable gradient propagation. Comprehensive evaluation using both synthetic and field datasets demonstrates GEB-MPU-Net’s superior performance compared to conventional time-frequency analysis methods and established networks, such as U-Net, residual dense network, residual dense block U-Net, and MPU-Net. The architecture consistently achieves enhanced reflection continuity, improved geological feature resolution, and robust noise suppression. These advancements provide more reliable input for seismic interpretation, better preservation of subtle stratigraphic features, and increased applicability to challenging field conditions.

Keywords

Noise suppression / Deep learning / Multi-stage networks / Seismic exploration

Funding

This work was supported in part by the Natural Science Foundation of Shanxi Province (202103021224012).

References

[1] Lee D, Shin RS, Yeo ME. Denoising sparker seismic data with deep BiLSTM in fractional Fourier transform. Comput Geosci. 2024;184:105519. doi:10.1016/j.cageo.2024.105519

[2] Geetha K, Kumar MH, Ajitha DK. A novel approach for seismic signal denoising using optimized discrete wavelet transform via honey badger optimization algorithm. J Appl Geophys. 2023;219:105236. doi:10.1016/j.jappgeo.2023.105236

[3] Zhang S, Zhang D, Yang X, Xu S, Sun Z, Liu Y. Noise reduction method based on curvelet theory of seismic data. Pet Sci Technol. 2023; 41(24):2344-2361. doi:10.1080/10916466.2022.2118771

[4] Li M, Li Y, Wu N, Wu J, Ma Y, Liu Q. Desert seismic data denoising based on energy spectrum analysis in empirical curvelet domain. Stud Geophys Geod. 2020; 64(2):373-390. doi:10.1007/s11200-019-0476-4

[5] Dalai B, Kumar P, Yuan X. De-noising receiver function data using the Seislet transform. Geophys J Int. 2019; 217(3):2047-2055. doi:10.1093/gji/ggz135

[6] Hasan MDA, Ahmad ZAB, Leong MS, Hee LM. Automated denoising technique for random input signals using empirical mode decomposition (EMD)-stabilization diagram. MATEC Web Conf. 2019;255:01004. doi:10.1051/matecconf/201925501004

[7] Liu W, Liu Y, Li S, Chen Y. A review of variational mode decomposition in seismic data analysis. Surv Geophys. 2023; 44(2):323-355. doi:10.1007/s10712-022-09742-z

[8] Zhao YX, Li Y, Yang BJ. Denoising of seismic data in desert environment based on a variational mode decomposition and a convolutional neural network. Geophys J Int. 2020; 221(2):1211-1225. doi:10.1093/gji/ggaa071

[9] Li J, Fan W, Li Y, Li W, Zhang Y, Zeng M. Desert seismic noise suppression based on an improved low-rank matrix approximation method. J Appl Geophys. 2020;173:103926. doi:10.1016/j.jappgeo.2019.103926

[10] Sun F, Zhang Q, Wang Z, Li Y. Compressed sensing with logsum heuristic recover for seismic denoising. Front Earth Sci. 2023;11:1285622. doi:10.3389/feart.2023.1285622

[11] Liu L, Ma J. Structured graph dictionary learning and application on the seismic denoising. IEEE Trans Geosci Remote Sens. 2019; 57(4):1883-1893. doi:10.1109/TGRS.2018.2870087

[12] Liu Q, Fu L, Zhang M. Deep-seismic-prior-based reconstruction of seismic data using convolutional neural networks. Geophysics. 2021; 86(2):V131-V142. doi:10.1190/geo2019-0570.1

[13] Li X, Hawari KG, Huang F, Liu J, Zhang Y. Review of CNN in aerial image processing. Imaging Sci J. 2023; 71(1):1-13. doi:10.1080/13682199.2023.2174651

[14] Wei D, Chen G, Chen J, Li C, Wu S, Zhang Y. Seismic data denoising using a self-supervised deep learning network. Math Geosci. 2023; 56(3):487-510. doi:10.1007/s11004-023-10089-3

[15] Li J, Qu R, Lu C. Multiple attention mechanisms-based convolutional neural network for desert seismic denoising. Pure Appl Geophys. 2023; 180(6):2135-2155. doi:10.1007/s00024-023-03255-5

[16] Ji G, Wang C. A denoising method for seismic data based on SVD and deep learning. Appl Sci. 2022; 12(24):12840. doi:10.3390/app122412840

[17] Guo Z, Zhu S, Chen J, Zhu W. Research on deep convolutional neural network time-frequency domain seismic signal denoising combined with residual dense blocks. Front Earth Sci. 2021;9:770748. doi:10.3389/feart.2021.681869

[18] Cai J, Wang L, Zheng J, Duan Z, Li L, Chen N. Denoising method for seismic co-band noise based on a U-Net network combined with a residual dense block. Appl Sci. 2023; 13(3):1324. doi:10.3390/app13031324

[19] Ding M, Zhou Y, Chi Y. Seismic signal denoising using Swin-Conv-UNet. J Appl Geophys. 2024;223:105355. doi:10.1016/j.jappgeo.2024.105355

[20] Zamir SW, Arora A, Khan S, Hayat M, Khan F, Shah M. Multi-stage progressive image restoration. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Nashville, TN, USA, 2021. p. 14821-14831. doi:10.1109/CVPR46437.2021.01458

[21] Saad MO, Ravasi M, Alkhalifah T. Self-supervised multistage deep learning network for seismic data denoising. Artif Intell Geophysics. 2025; 6(1):100123. doi:10.1016/j.aiig.2025.100123

[22] Obou.1016/j.aiig.2025.100123123ing. learning network for seismic data denoising. enoising. ng. sing. ing. ng. l denoisingGeophysics. 2025;90(3):V205-V219. doi:10.1190/geo2024-0456.1

[23] Tian C, Yuan Y, Zhang S, Wang X, Li H. Image super-resolution with an enhanced group convolutional neural network. Neural Netw. 2022; 153:373-385. doi:10.1016/j.neunet.2022.06.009

[24] Saad OM, Ravasi M, Alkhalifah T. Noise attenuation in distributed acoustic sensing data using a guided unsupervised deep learning network. Geophysics. 2024; 89(6):V573-V587. doi:10.1190/geo2024-0109.1

[25] Saad OM, Obou23-0642Bai M, Alkhalifah T, Wang Y, Liu J. Self-attention deep image prior network for unsupervised 3-D seismic data enhancement. IEEE Trans Geosci Remote Sens. 2021; 60(5):1-14. doi:10.1109/TGRS.2021.3108515