Cite this article
1
Download
57
Views
Journal Browser
Volume | Year
Issue
Search
News and Announcements
View All
ARTICLE

Convolutional autoencoder neural network for seismic noise reduction

DARAPUREDDY HARITHA NITTALA SATYAVANI*
Show Less
CSIR - National Geophysical Research Institute, Hyderabad, India.,
JSE 2022, 31(3), 267–278;
Submitted: 28 June 2021 | Accepted: 14 February 2022 | Published: 1 June 2022
© 2022 by the Authors. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution -Noncommercial 4.0 International License (CC-by the license) ( https://creativecommons.org/licenses/by-nc/4.0/ )
Abstract

Haritha, D. and Satyavani, N., 2022. Convolutional autoencoder neural network for seismic noise reduction. Journal of Seismic Exploration, 31: 267-278. Seismic noise reduction is one of the main steps in the data processing sequence that aids in proper seismic imaging and interpretation. For the purpose of noise reduction and to recover weaker / masked signals, we propose the scheme of an unsupervised convolutional autoencoder neural network. Cross-entropy is used as the loss function in the network. The adaptive moment estimation plays the role of a backpropagation algorithm that can optimize the loss function. The key parameters of the network, like convolutional layers, filter size, and learning rate have been selected after performing a series of tests with different values for each of the parameters and those results are also presented here. We show that the present network applied to the seismic data shows improvement in the reflections and also allows us to recover some of the weaker / masked signals. The results show that the proposed method is successful in suppressing the noise and enhancing the seismic signals.

Keywords
seismic data
noise reduction
convolutional layers
filter size
convolutional autoencoder
References
  1. Bednar, J.B., 1983. Applications of median filtering to deconvolution, pulse estimation,and statistical editing of seismic data. Geophysics, 48: 1598-1610.
  2. Bekara, M. and van der Baan, M., 2007. Local singular value decomposition for signalenhancement of seismic data. Geophysics, 72(2): V59-V65.
  3. Burger, H.C., Schuler, C.J. and Harmeling, S., 2012. Image denoising. Can plain neuralnetworks compete with BM3D? IEEE Conf. Comput. Vis. Pattern Recognit.(CVPR), 2392-2399.
  4. Chen, H., Zhang, Y., Kalra, M.K., Lin, F., Chen, Y., Liao, P., Zhou, J. and Wang, G.,
  5. Low-dose ct with a residual encoder-decoder convolutional neural network.IEEE Transact. Medic. Imag., 36: 2524-2535.
  6. Chen, Y. and Pock, T., 2017. Trainable nonlinear reaction-diffusion: A flexibleframework for fast and effective image restoration. IEEE Transact. PatternAnalys. Mach. Intellig., 39: 1256-1272.
  7. Chen, Y., Zhang, M., Bai, M. and Chen, W., 2019. Improving the signal-to-noise ratio ofseismological datasets by unsupervised machine learning. Seismol. Res. Lett., 90:1552-1564.
  8. Du, B., Xiong, W., Wu, J., Zhang, L., Zhang, L. and Tao, D., 2016. Stackedconvolutional denoising auto-encoders for feature representation. IEEE Transact.Cybernet., 47: 1017-1027.
  9. Duncan, G. and Beresford, G., 1995. Some analyses of 2-D median f-k filters.Geophysics, 60: 1157-1168.
  10. Glinsky, M.E., Clark, G.A. and Cheng, P.K.Z., 1996. Automatic event picking inprestack migrated gathers using a probabilistic neural network. Geophysics, 66:1488-1496.
  11. Gondara, L., 2016. Medical image denoising using convolutional denoisingautoencoders. 16th Internat. IEEE Conf. Data Mining Workshops (ICDMW):241-246.
  12. Harris, P. and White, R., 1997. Improving the performance of f-x prediction filtering atlow signal-to-noise ratios. Geophys. Prosp., 45: 269-302.
  13. Huang, W., Wang, R., Chen, X. and Chen, Y., 2017a. Double least-squares projectionsmethod for signal estimation. IEEE Transact. Geosci. Remote Sens. Lett., 55:4111-4129.
  14. Huang, L., Dong, X. and Clee, T.E., 2017. A scalable deep learning platform foridentifying geologic features from seismic attributes. The Leading Edge, 36(3):249-256.
  15. Jain, V. and Seung, S., 2009. Natural image denoising with convolutional networks.
  16. Advan. Neur. Informat. Process. Syst., (Vol. No. ???) 769-776.
  17. Kingma, D.P. and Ba, 本, 2014. Adam: a method for stochastic optimization. CoRR., Vol.abs/ 1412.6980, 1-15.
  18. Kline, D.M. and Berardi, V.L., 2005. Revisiting squared-error and cross-entropyfunctions for training neural network classifiers. Neural Comput. Applicat., 14:310-318.
  19. Kononenko, I, 2001. Machine learning for medical diagnosis: history, state of the artand perspective. Artific. Intellig. Medic., 23: 89-109.
  20. Krizhevsky, A., Sutskever, I. and Hinton, G.E., 2017. ImageNet classification with deepconvolutional neural networks. Communicat. ACM, 60(6): 84-90.
  21. Li, T. and Mallick, S., 2014. Multicomponent, multi-azimuth pre-stack seismicwaveform inversion for azimuthally anisotropic media using a parallel andcomputationally efficient non-dominated sorting genetic algorithm. Geophys. J.Internat., 200: 1136-1154.
  22. Nath, S.K., Chakraborty, S., Singh, S.K. and Ganguly, N., 1999. Velocity inversion incross-hole seismic tomography by a counter-propagation neural network, geneticalgorithm, and evolutionary programming techniques. Geophys. J. Internat., 138:108-124.
  23. Pratt, R.G., 1999. Seismic waveform inversion in the frequency domain, Part 1: Theoryand verification in a physical scale model. Geophysics, 64: 888-901.
  24. Ross, C.P. and Cole, D.M., 2017. A comparison of popular neural network facies-classification schemes. The Leading Edge, 36(4): 340-349.
  25. Rowley, H.A., Baluja, S. and Kanade, T., 1998. Neural network-based face detection.
  26. IEEE Transact. Patt. Analys. Mach. Intellig., 20: 23-38.
  27. Satyavani, N., Sain, K., Lall, M. and Kumar, B.J.P., 2008. Seismic attribute study for gashydrates in the Andaman Offshore India. Marine Geophys. Res., 29: 167-175.
  28. Song, H., Yang, G., Chen, W., Xue, Y., Zhang, H. and Zhang, X., 2020. Seismic randomnoise suppression using deep convolutional autoencoder neural network. J. Appl.Geophys., 178: 104071.
  29. Wang, Y., Zhou, H., Zu, S., Mao, W. and Chen, Y., 2017. Three-operator proximalsplitting scheme for 3D seismic data reconstruction. IEEE Geosci. Remote Sens.Lett., 14: 1830-1834.
  30. Wu, Z. and Liu, T., 2008. Wavelet transform methods in seismic data noise attenuation.Progr. Geophys., 23: 493- 499.
  31. Zhang, K., Zuo, W., Chen, Y., Meng, D. and Zhang, L., 2017. Beyond a Gaussiandenoiser. Residual learning of deep CNN for image denoising. IEEE Transact.Image Process., 26: 3142-3155.
  32. Zhang, M., Liu, Y., Bai, M. and Chen, Y., 2019. Seismic noise attenuation usingunsupervised sparse feature learning. IEEE Transact. Geosci. Remote Sens.Lett., 57: 970-972.
  33. Zhao, Q., Du, Q., Sun, W. and Chen, Y., 2019. Iterative double Laplacian-scaled low-rank optimization for under-sampled and noisy signal recovery. Expanded Abstr.,89th Ann. Internat. SEG Mtg., San Antonio: 4675-4679.
Share
Back to top
Journal of Seismic Exploration, Electronic ISSN: 0963-0651 Print ISSN: 0963-0651, Published by AccScience Publishing