Attention mechanism-based deep denoiser for desert seismic random noise suppression

• Autorzy: Lin, Hongbo , Liu, Chang , Wang, Shigang , Ye, Wenhai
• Języki publikacji: EN

Abstrakty

EN

Seismic data collected from desert areas contain a large amount of low-frequency random noise with similar waveforms to the effective signals. The complex noise characteristics make it difficult to effectively identify and recover seismic signals, which will adversely affect subsequent seismic data processing and imaging. In order to recover the complex seismic events from low-frequency random noise, we propose an attention mechanism guided deep convolutional autoencoder network (ADCAE) to assign different importance to different features at different spatial position. In ADCAE, an attention module (AM) is connected to the deep convolutional autoencoder network (DCAE) with soft-thresholded symmetric skip connection that helps to enhance the ability of feature extraction. By combining the global features of the input data and the output local features of DCAE, AM generates an attention weight matrix, which assigns different weights to the features associated with the seismic events and random noise during the training process. In this way, AM can guide the update of the target gradient, thus retains the complex structure of the seismic events in the denoised results and improves the training efficiency of the model. The ADCAE is applied to the synthetic data and field seismic data, and denoised results show that ADCAE has achieved satisfactory denoising performance in signals recovery and low-frequency random noise suppression at the low signal-to-noise ratio.

Słowa kluczowe

EN

seismic explorations; random noise suppression; deep convolutional autoencoder network; attention mechanism; low signal-to-noise ratio

• Czasopismo: Acta Geophysica
• Rocznik: 2023
• Tom: Vol. 71, no. 6
• Strony: 2781--2793
• Opis fizyczny: Bibliogr. 34 poz., rys., tab.

Twórcy

autor

Lin, Hongbo

• College of Communication and Engineering, Jilin University, Changchun, 130012, China,

autor

Liu, Chang

• College of Communication and Engineering, Jilin University, Changchun, 130012, China,

autor

Wang, Shigang

• College of Communication and Engineering, Jilin University, Changchun, 130012, China

autor

Ye, Wenhai

• Jilin Province Kewei Traffic Engineering Co., Ltd, Changchun, 130000, China

Bibliografia

• 1. Aaditya P, James S, Dinei F, et al (2019) RePr: improved training of convolutional filters. In: 2019 IEEE/cvf conference on computer vision and pattern recognition USA
• 2. Chen Y, Fomel S (2015) Random noise attenuation using local signal-and-noise orthogonalization. Geophysics 80(6):WD1–WD9
• 3. Feng Q, Li Y, Wang H (2020) Intelligent random noise modeling by the improved variational autoencoding method and its application to data augmentation. Geophysics 86(1):19–31. https://doi.org/10.1190/geo2019-0815.1
• 4. Hu J, Shen L, Albanie S et al (2020) Squeeze-and-excitation networks. IEEE Trans Pattern Anal Mach Intell 42(8):2011–2023. https://doi.org/10.1109/TPAMI.2019.2913372
• 5. Krizhevsky A, Sutskever I, Hinton G (2012) ImageNet classification with deep convolutional neural networks. Adv Neural Inf Process Syst 25(2):84–90. https://doi.org/10.1145/3065386
• 6. Li G, Li Y, Yang B (2017) Seismic exploration random noise on land: modeling and application to noise suppression. IEEE Trans Geosci Remote Sens 55(8):4668–4681. https://doi.org/10.1109/TGRS.2017.2697444
• 7. Li Y, Wang H, Dong X (2021) The denoising of desert seismic data based on cycle-gan with unpaired data training. IEEE Geosci Remote Sens Lett 18(11):2016–2020. https://doi.org/10.1109/LGRS.2020.3011130
• 8. Lin H, Wang S, Li Y (2021) A branch construction-based CNN denoiser for desert seismic data. IEEE Geosci Remote Sens Lett 18(4):736–740. https://doi.org/10.1109/LGRS.2020.2981965
• 9. Ma H, Qian Z, Li Y et al (2019a) Noise reduction for desert seismic data using spectral kurtosis adaptive bandpass filter. Acta Geophys 67(1):123–131. https://doi.org/10.1007/s11600-018-0232-0
• 10. Ma H, Yao H, Li Y et al (2019b) Deep residual encoder-decoder networks for desert seismic noise suppression. IEEE Geosci Remote Sens Lett 17(3):529–533. https://doi.org/10.1109/LGRS.2019.2925062
• 11. Ma H, Yan J, Li Y et al (2019c) Desert seismic random noise reduction based on LDA effective signal detection. Acta Geophys 67:109–121. https://doi.org/10.1007/s11600-019-00250-0
• 12. Mao X, Shen C, Yang Y (2016) Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections. In: 30th conference on neural information processing systems (NIPS)
• 13. Masci J, Meier U, Dan C et al (2011) Stacked convolutional auto-encoders for hierarchical feature extraction. Lect Notes Comput Sci 6791(1):52–59. https://doi.org/10.1007/978-3-642-21735-7_7
• 14. Niu R, Sun X, Tian Y et al (2022) Hybrid multiple attention network for semantic segmentation in aerial images. IEEE Trans Geosci Remote Sens 60:1–18. https://doi.org/10.1109/TGRS.2021.3065112
• 15. Rumelhart DE, Hinton GE, Williams RJ (1986) Learning representations by back propagating errors. Nature 323(6088):533–536. https://doi.org/10.1038/323533a0
• 16. Saad O, Chen Y (2021) A fully unsupervised and highly generalized deep learning approach for random noise suppression. Geophys Prospect 69(4):709–726. https://doi.org/10.1111/1365-2478.13062
• 17. Saad O, Bai M, Chen Y (2021) Uncovering the microseismic signals from noisy data for high-fidelity 3D source-location imaging using deep learning. Geophysics 60(6):ks161–ks173. https://doi.org/10.1190/GEO2021-0021.1
• 18. Sang W, Yuan S, Yong X et al (2021) DCNNs-based denoising with a novel data generation for multidimensional geological structures learning. IEEE Geosci Remote Sens Lett 18(10):1861–1865. https://doi.org/10.1109/LGRS.2020.3007819
• 19. Sang W, Yuan S, Han H et al (2023) Porosity prediction using semi-supervised learning with biased well log data for improving estimation accuracy and reducing prediction uncertainty. Geophys J Int 232(2):940–957. https://doi.org/10.1093/gji/ggac371
• 20. Song H, Gao Y, Chen W et al (2020) Seismic random noise suppression using deep convolutional autoencoder neural network. J Appl Geophys 178:0926–9851. https://doi.org/10.1016/j.jappgeo.2020.104071
• 21. Sun X, Li Y (2020) Denoising of desert seismic signal based on synchrosqueezing transform and Adaboost algorithm. Acta Geophys 68:403–412. https://doi.org/10.1007/s11600-020-00408-1
• 22. Vincent P, Larochelle H, Bengio Y et al (2008) Extracting and composing robust features with denoising autoencoders. In: Proceedings of the twenty-fifth international conference
• 23. Wang S, Li Y, Zhao Y (2020) Desert seismic noise suppression based on multimodal residual convolutional neural network. Acta Geophys 68(2):389–401. https://doi.org/10.1007/s11600-020-00405-4
• 24. Wilterson A, Graziano M (2021) The attention schema theory in a neural network agent: controlling visuospatial attention using a descriptive model of attention. Proc Natl Acad Sci. https://doi.org/10.1073/pnas.2102421118
• 25. Woo S, Park J, Lee J et al (2018) CBAM: convolutional block attention module. In: Computer vision. 15th European conference
• 26. Yang H, Chen Y, Song K et al (2019) Multiscale feature – clustering—based fully convolutional autoencoder for fast accurate visual inspection of texture surface defects. IEEE Trans Autom Sci Eng 16(3):1450–1467. https://doi.org/10.1109/TASE.2018.2886031
• 27. Yang H, Kim J, Kim H et al (2020) Guided soft attention network for classification of breast cancer histopathology images. IEEE Trans Med Imaging 39(5):1306–1315. https://doi.org/10.1109/TMI.2019.2948026
• 28. Yang L, Wang S, Chen X et al (2022) Unsupervised 3-D random noise attenuation using deep skip autoencoder. IEEE Trans Geosci Remote Sens 60:5905416. https://doi.org/10.1109/TGRS.2021.3100455
• 29. Yuan S, Jiao X, Luo Y et al (2021) Double-scale supervised inversion with a data-driven forward model for low-frequency impedance recovery. Geophysics 87(2):R165–R181. https://doi.org/10.1190/geo2020-0421.1
• 30. Zhao B, Wu X, Peng Q et al (2017) Diversified visual attention networks for fine-grained object classification. IEEE Trans Multimedia 19(6):1245–1256. https://doi.org/10.1109/TMM.2017.2648498
• 31. Zhao Y, Li Y, Dong X et al (2019) Low-frequency noise suppression method based on improved DnCNN in desert seismic data. IEEE Geosci Remote Sens Lett 16(5):811–815. https://doi.org/10.1109/LGRS.2018.2882058
• 32. Zhao Y, Li Y, Yang B (2020a) Low-frequency desert noise intelligent suppression in seismic data based on multiscale geometric analysis convolutional neural network. IEEE Trans Geosci Remote Sens 58(1):650–665. https://doi.org/10.1109/TGRS.2019.2938836
• 33. Zhao Q, Du Q, Li Q et al (2020) Robust dictionary learning for erratic noise-corrupted seismic data reconstruction. Acta Geophys 68:687–700. https://doi.org/10.1007/s11600-020-00433-0
• 34. Zhang Y, Lin H, Li Y et al (2020) Seismic signal enhancement and noise suppression using structure-adaptive nonlinear complex diffusion. IEEE Trans Geosci Remote Sens 58(3):2198–2211. https://doi.org/10.1109/TGRS.2019.2954949

Uwagi

Opracowanie rekordu ze środków MNiSW, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2024).

Identyfikatory

DOI

10.1007/s11600-023-01062-z