Hierarchical division encoder–decoder network for distributed acoustic sensing-vertical seismic profile noise suppression

Abstract

In recent years, traditional geophones for well seismic data acquisition have progressively been replaced by distributed acoustic sensing (DAS), a novel technique. The primary attributes of DAS are its extensive well coverage and robust adaptability to challenging acquisition situations. Unlike conventional geophones, vertical seismic profile (VSP) data obtained using DAS exhibit lower signal-to-noise ratios (SNRs) and more complex noise types. These complex and energetic perturbations pose challenges for further data analysis. Contemporary methods for mitigating noise in DAS-VSP data sometimes fail to yield complete and precise information, leading to inferior denoising quality and diminished signal recovery. We propose a hierarchical division encoder–decoder network utilizing a convolutional neural network to address this issue. This network employs spatial attention techniques for systematic reconstruction and facilitates hierarchical feature extraction according to information scale. Our methodology provides superior noise reduction capabilities while preserving signal integrity. It achieves this by comprehensively addressing features at all scales. Additionally, we generated the required training set by combining synthetic data with real noise, as no publicly available training sets are available for DAS-VSP data. The trained denoising network processes and analyzes both synthetic and real recordings. The experimental results demonstrate the efficacy of this technique in eliminating various types of DAS-VSP noise while preserving signal amplitude integrity and ensuring continuity of signal recovery.

Keywords

Hierarchical division / Noise suppression / Convolutional neural network / Signal-to-noise ratio / Attention mechanism

References

[1] Wang Q, Zhang Y, Xie Z, et al. The Advancement and Challenges of Seismic Techniques for Ultra—Deep Carbonate Reservoir Exploitation in the Tarim Basin of Northwestern China. Energies. 2022; 15(20): 7653. doi: 10.3390/en15207653

[2] Dong XT, Li Y, Yang BJ. Desert low—frequency noise suppression by using adaptive DnCNNs based on the determination of high—order statistic. Geophys J Int. 2019; 219(2): 1281-1299. doi: 10.1093/gji/ggz363

[3] Wang Z, Fan Z, Chen X, et al. Global oil and gas development situation, trends and enlightenment in 2023. Pet Explor Dev . 2024; 51(6): 1536-1555. doi: 10.1016/S1876—3804(25)60558—1

[4] Zhu Q, Huang H, Wei X, et al. Fu Za You Qi Cang Chu Ceng Heng Xiang Yu Ce [Lateral prediction of complex oil and gas reservoirs]. Oil Geophysical Prospecting. 2003; 38(004): 405-408. [In Chinese] doi: 10.3321/j.issn:1000—7210.2003.04.013

[5] Wang Z, Lu B, Ye Q, Cai H. Recent Progress in Distributed Fiber Acoustic Sensing with Φ—OTDR. Sensors. 2020; 20(22): 6594. doi: 10.3390/s20226594

[6] Rashid A, Tackie—Otoo BN, Abdul Latiff AH, et al. Research Advances on Distributed Acoustic Sensing Technology for Seismology. Photonics. 2025; 12(3): 196. doi: 10.3390/photonics12030196

[7] Liu Z, Gong H, Wu Y, et al. Xiang Gan Guang Shi Yu Fan She Yi Ji Shu Yu Ying Yong [Principle and Application of the C—OTDR]. Study Opt Commun. 2022; 48(5): 43— 48, 65. [In Chinese] doi: 10.13756/j.gtxyj.2022.05.008

[8] Wang L, Xie X, Yao M, et al. Fen Bu Shi Guang Xian Sheng Bo Chuan Gan Ji Qi Jin Zhan Yan Jiu [Distributed Optical Fiber Acoustic Wave Sensing and Its Progress]. Study Opt Commun. 2025; 51(5): 240100—01. [In Chinese] doi: 10.13756/j.gtxyj.2025.240100

[9] Zhang Z, Alajami M, Alkhalifah T. Wave—equation dispersion spectrum inversion for near—surface characterization using fiber—optics acquisition. Geophys J Int. 2020; 222(2): 907-918. doi: 10.1093/gji/ggaa211

[10] Isaenkov R, Tertyshnikov K, Yurikov A, et al. Effect of Source Mispositioning on the Repeatability of 4D Vertical Seismic Profiling Acquired with Distributed Acoustic Sensors. Sensors. 2022; 22(24): 9742. doi: 10.3390/s22249742

[11] Mad Zahir MH, Abdul Aziz KM, Ghazali AR, et al. Effectiveness of Fiber Optic Distributed Acoustic Sensing (DAS) in Vertical Seismic Profiling (VSP) Field Survey. Appl Sci. 2023; 13(8): 5002. doi: 10.3390/app13085002

[12] Zhang S, Wang J, Yang Z. Walkaway VSP Chu Li Ji Shu Ji Ying Yong [The processing technology and application of The Walkaway VSP]. Proc SPIE. 2017; 52(2): 1-6. [In Chinese] doi: 10.13810/j.cnki.issn.1000—7210.2017.S2.001

[13] Lin RB, Zeng XF, Song ZH, et al. Fen Bu Shi Guang Xian Sheng Bo Chuan Gan Xi Tong Zai Jin Di Biao Cheng Xiang Zhong De Ying Yong 2: Bei Jing Zao Sheng Cheng Xiang [Distributed acoustic sensing for imaging shallow structure Ⅱ: Ambient noise tomography]. Chinese J Geophys. 2020; 63(4): 1622-1629. [In Chinese] doi: 10.6038/cjg2020N0272

[14] Saad OM, Ravasi M, Alkhalifah T. Noise attenuation in distributed acoustic sensing data using a guided unsupervised deep learning network. Geophysics. 2025; 89(6): V573-V587. doi: 10.1190/GEO2024—0109.1

[15] Yu G, Chen YZ, Wu J, et al. 3D—VSP Survey Using a DAS System and Downhole Geophone Array in Southwest China. In: Proceedings of 81st EAGE Conference and Exhibition. London, UK; 2019: 1-5. doi: 10.3997/2214—4609.201901252

[16] Yu G, Xiong JL, Wu JJ, et al. Enhanced surface seismic data processing using simultaneous acquired DAS—VSP data. First Break . 2020; 38(6): 29-36. doi: 10.3997/1365—2397.fb2020039

[17] Chavarria A, Minto C, Karrenbach M, et al. Offshore DASVSP Applications and Long—Range Distributed Fiber Optic Sensing. In: First EAGE Workshop on Fibre Optic Sensing . European Association of Geoscientists & Engineers; 2020; 2020: 1-5. doi: 10.3997/2214—4609.202030018

[18] Zhu D, Fu L, Kazei V, Li W. Diffusion Model for DAS—VSP Data Denoising. Sensors. 2023; 23(20): 8619. doi: 10.3390/s23208619

[19] Zhong T, Cheng M, Dong X, Li Y, Wu N. Seismic random noise suppression by using deep residual U—Net. J Petrol Sci Eng. 2022; 209: 109901. doi: 10.1016/j.petrol.2021.109901

[20] Gu X, Collet O, Tertyshnikov K, Pevzner R. Removing Instrumental Noise in Distributed Acoustic Sensing Data: A Comparison Between Two Deep Learning Approaches. Remote Sens. 2024; 16(22): 4150. doi: 10.3390/rs16224150

[21] Harsuko R, Alkhalifah T. Optimizing a transformer—based network for a deep—learning seismic processing workflow. Geophysics. 2024; 89(4): V347-V359. doi: 10.1190/GEO2023—0403.1

[22] Lou Y, Wang Y, Zhang X, et al. Seismic Reflector Dip Integrated Spatial—Spectral UFormer for Fault Detection. IEEE Trans Geosci Remote Sens. 2025; 63: 1-15. doi: 10.1109/TGRS.2025.3526814

[23] Lou Y, Zhang X, Wang Y, et al. Unconformity—Enhanced Hierarchical Context Fusion Network for Seismic Facies Classification. IEEE Trans Geosci Remote Sens. 2025; 63: 1-12. doi: 10.1109/TGRS.2025.3563362

[24] Lou Y, Wang Y, Chen Y, et al. Automatic Seismic Fault Interpretation With the Aid of Data—, Physics—, and Math—Assisted Synthetic Data Generation. IEEE Trans Geosci Remote Sens. 2024; 62: 1-15. doi: 10.1109/TGRS.2024.3402739

[25] Zhou X, Qiao L. Di Zhen Kan Tan Zhong De Chang Jian Di Zhen Gan Rao Bo Ji Ya Zhi Fang Fa [Common seismic interference waves and suppression methods in seismic exploration]. West China Explor Eng. 2008; 20(11): 138-141. [In Chinese] doi: 10.3969/j.issn.1004—5716.2008.11.048

[26] Tang S, Lin J. Xiao Bo Bian Huan Yu Zhong Zhi Lü Bo Xiang Jie He Tu Xiang Qu Zao Fang Fa [Image denoising with combination of wavelet transform and median filtering]. J Harbin Inst Technol. 2008; 40(8): 1334-1336. [In Chinese] doi: 10.3321/j.issn:0367—6234.2008.08.034

[27] Bekara M, van der Baan M. Random and coherent noise attenuation by empirical mode decomposition. Geophysics . 2009; 74(5): V89-V98. doi: 10.1190/1.3157244

[28] Wang W, Yang G. Ji Yu EMD Yu ICA De Di Zhen Xin Hao Qu Zao Ji Shu Yan Jiu [Research on Seismic Signal Denoising Technology Based on EMD and ICA]. Geophys Prospect Pet. 2012; 51(1): 19-29. [In Chinese] doi: 10.3969/j.issn.1000—1441.2012.01.003

[29] Yao Z, Li W, Zhu J, Hao L, Xing M. Seismic random noise separation and suppression based on improved variational mode decomposition via grey wolf optimization. PLoS ONE. 2025; 20(9): e0330988. doi: 10.1371/journal.pone.0330988

[30] Nie R, Xu H, Zhao Y. Lian He VMD He Shearlet Bian Huan De Qu Zao Fang Fa Yan Jiu [Research on Denoising Methods by Combining VMD and Shearlet Transform]. Chin J Eng Geophys. 2024; 21(1): 176-186. [In Chinese] doi: 10.3969/j.issn.1672—7940.2024.01.018

[31] Han J, van der Baan M. Microseismic and seismic denoising via ensemble empirical mode decomposition and adaptive thresholding. Geophysics . 2015; 80(6): KS69-KS80. doi: 10.1190/GEO2014—0423.1

[32] Li X, Yang K, Liu W, et al. Ji Yu Gai Jin De EEMD Shi Xian Di Zhen Xin Hao Qu Zao [Seismic Signal De—noising Based on Improved EEMD]. Chin J Eng Geophys. 2014; 11(4): 431-435. [In Chinese] doi: 10.3969/j.issn.1672—7940.2014.04.001

[33] Tang HH, Cheng SJ, Li WQ, et al. An EDCC—EMD analysis—based network for DAS VSP data denoising in frequency domain. Pet Sci. 2025; 22(5): 1929-1945. doi: 10.1016/j.petsci.2025.03.006

[34] Kong Q, Wu F, Fan B. Ji Yu Shen Du Xue Xi De Tu Xiang Pi Pei: Fang Fa Ying Yong Yu Tiao Zhan [Image Matching in Deep Learning Era: Methods, Applications and Challenges]. Chin J Comput. 2024; 47(7): 1485-1520. [In Chinese] doi: 10.11897/SP.J.1016.2024.01485

[35] Li H, Liu F, Yang SY, et al. Ji Yu Shen Du Zhi Cheng Zhi Xue Xi Wang Luo De Yao Gan Tu Xiang Rong He [Remote sensing image fusion based on deep support value learning networks]. Chin J Comput. 2016; 39(8): 1583-1596. [In Chinese] doi: 10.11897/SP.J.1016.2016.01583

[36] Lou Y, Zhang X, Liu N, et al. Seismic Unconformity Estimation via ECA—UNet plus plus With Physical Knowledge Constrained Synthetic Data. IEEE Trans Geosci Remote Sens. 2024; 62: 5933414. doi: 10.1109/TGRS.2024.3481448

[37] Saad OM, Ravasi M, Alkhalifah T. Self—supervised multi—stage deep learning network for seismic data denoising. Artif Intell Geosci. 2025; 6(1): 100123. doi: 10.1016/j.aiig.2025.100123

[38] Han W, Zhou Y, Chi Y. Ji Yu Shen Du Xue Xi Juan Ji Shen Jing Wang Luo De Di Zhen Shu Ju Sui Ji Zao Sheng Qu Chu [Deep learning convolutional neural networks for random noise attenuation in seismic data]. Geophys Prospect Pet. 2018; 57(6): 862—869, 877. [In Chinese] doi: 10.3969/j.issn.1000—1441.2018.06.008

[39] Liu S, Birnie C, Alkhalifah T. Trace—wise coherent noise suppression via a self—supervised blind—trace deep—learning scheme. Geophysics. 2023; 88(6): V459-V472. doi: 10.1190/GEO2022—0371.1

[40] Wang Y, Lou Y, Liu N, et al. Seismic Fault Delineation and Parameter Interpretation Using Dual U—Shaped Domain Fusion—Net. IEEE Trans Geosci Remote Sens. 2025; 63: 1-15. doi: 10.1109/TGRS.2025.3612429

[41] Izzatullah M, Alkhalifah T, Romero J. Posterior sampling with convolutional neural network—based plug—and—play regularization with applications to poststack seismic inversion. Geophysics. 2024; 89(2): R137-R153. doi: 10.1190/geo2023—0035.1

[42] Zhang K, Zuo W, Chen Y, et al. Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising. IEEE Trans Image Process. 2017; 26(7): 3142-3155. doi: 10.1109/TIP.2017.2662206

[43] Lehtinen J, Munkberg J, Hasselgren J, et al. Noise2noise: Learning image restoration without clean data. arXiv. 2018. doi: 10.48550/arXiv.1803.04189

[44] Abedi MM, Pardo D, Alkhalifah T. Semi—blind—trace algorithm for self—supervised attenuation of trace—wise coherent noise. Geophys Prospect. 2024; 72(3): 965-977. doi: 10.1111/1365—2478.13448

[45] Ding X, Zhang X, Han J, et al. Scaling Up Your Kernels to 31×31: Revisiting Large Kernel Design in CNNs. In: Proceedings of 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New Orleans, LA, USA: IEEE; 2022: 11953-11965. doi: 10.1109/CVPR52688.2022.01166

[46] Liu JW, Liu JW, Luo XL. Shen Du Xue Xi Zhong Zhu Yi Li Ji Zhi Yan Jiu Jin Zhan [Research progress in attention mechanism in deep learning]. Chin J Eng. 2021; 43(11): 1499-1511. [In Chinese] doi: 10.13374/j.issn2095—9389.2021.01.30.005

[47] Woo S, Park J, Lee JY, Kweon IS. CBAM: Convolutional block attention module. In: Ferrari V, Hebert M, Sminchisescu C, Weiss Y, (eds). Computer Vision — ECCV 2018; ECCV 2018; Lecture Notes in Computer Science (LNIP, volume 11211). Springer International Publishing; 2018; 11211: 3-19. doi: 10.1007/978—3—030—01234—2_1

[48] Ren H, Wang X. Zhu Yi Li Ji Zhi Zong Shu [Review of attention mechanism]. J Comput Appl. 2021; 41(S1): 1-6. [In Chinese] Available from: https://cnki.com.cn/Article/CJFDTotal—JSJY2021S1002.html

[49] Hu J, Shen L, Sun G. Squeeze—and—excitation networks. In: Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE; 2018: 7132-7141. doi: 10.1109/CVPR.2018.00745

[50] Chen Y, Li W, Sakaridis C, et al. Domain Adaptive Faster R—CNN for Object Detection in the Wild. In: Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE; 2018: 3339-3348. doi: 10.1109/CVPR.2018.00352

[51] Liu B, Yue J, Zuo Z, et al. Unsupervised deep learning for random noise attenuation of seismic data. IEEE Geosci Remote Sens Lett. 2022; 19: 1-5. doi: 10.1109/LGRS.2021.3057631

[52] Roncoroni G, Fortini C, Bortolussi L, Bienati N, Pipan M. Synthetic seismic data generation with deep learning. J Appl Geophy. 2021; 190: 104347. doi: 10.1016/j.jappgeo.2021.104347