Abstract
Based on the CNN-LSTM fusion deep neural network, this paper proposes a seismic velocity model building method that can simultaneously estimate the root mean square (RMS) velocity and interval velocity from the common-midpoint (CMP) gather. In the proposed method, a convolutional neural network (CNN) Encoder and two long short-term memory networks (LSTMs) are used to extract spatial and temporal features from seismic signals, respectively, and a CNN Decoder is used to recover RMS velocity and interval velocity of underground media from various feature vectors. To address the problems of unstable gradients and easily fall into a local minimum in the deep neural network training process, we propose to use Kaiming normal initialization with zero negative slopes of rectified units and to adjust the network learning process by optimizing the mean square error (MSE) loss function with the introduction of a freezing factor. The experiments on testing dataset show that CNN-LSTM fusion deep neural network can predict RMS velocity as well as interval velocity more accurately, and its inversion accuracy is superior to that of single neural network models. The predictions on the complex structures and Marmousi model are consistent with the true velocity variation trends, and the predictions on field data can effectively correct the phase axis, improve the lateral continuity of phase axis and quality of stack section, indicating the effectiveness and decent generalization capability of the proposed method.
Article PDF
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
References
Alzahrani, H., and Shragge, J., 2021, Neural network seismic velocity model building: A frequency-stepping approach: First International Meeting for Applied Geoscience & Energy, Society of Exploration Geophysicists, 3370–3374.
Biswas, R., Arnulf, A. F., Sen, M. K., et al, 2020, Two-step velocity inversion using trans-dimensional tomography and elastic FWI: 90th Annual International Meeting, SEG, Expanded Abstracts, 3628–3633.
Bouvrie, J., 2006, Notes on convolution neural networks. [Online]. Available: http://cogprints.org/5869/1/cnn_tutorial.pdf
Chai, X., Tang, G., Peng, R., et al., 2018, The linearized bregman method for frugal full-waveform inversion with compressive sensing and sparsity-promoting: Pure and Applied Geophysics, 175(3), 1085–1101.
Dix, C. H., 1955, Seismic velocities from surface measurements: Geophysics, 20(1), 68–86.
Fabien-Ouellet, G., and Sarkar, R., 2020, Seismic velocity estimation: A deep recurrent neural-network approach: Geophysics, 85(1), U21–U29.
Girshick, R., 2015, Fast R-CNN: 2015 IEEE International Conference on Computer Vision, 1440–1448.
Glorot, X., and Bengio, Y., 2010, Understanding the difficulty of training deep feedforward neural networks: Proceedings of the 13th international conference on artificial intelligence and statistics, 249–256.
Graves, A., Eck, D., Beringer, N., et al., 2004, Biologically plausible speech recognition with LSTM neural nets: International Workshop on Biologically Inspired Approaches to Advanced Information Technology, Springer, Berlin, Heidelberg, 127–136.
Graves, A., Mohamed, A. R., and Hinton, G., 2013, Speech recognition with deep recurrent neural networks: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 6645–6649.
Guitton, A., 2012, Blocky regularization schemes for full-waveform inversion: Geophysical Prospecting, 60(5), 870–884.
Gulordava, K., Bojanowski, P., Grave, E., et al., 2018, Colorless green recurrent networks dream hierarchically: arXiv: 1803.11138. [Online]. Available: https://arxiv.org/abs/1803.11138
Guo, X., Shi, Y., Wang, W., et al., 2019, A robust source-independent misfit function for time domain waveform inversion based on normalized convolved wavefield: Journal of Applied Geophysics, 166, 129–146.
Han, J., and Moraga, C., 1995, The influence of the sigmoid function parameters on the speed of backpropagation learning: International Workshop on Artificial Neural, Networks (pp. 195–201). Springer, Berlin, Heidelberg.
He, K., Zhang, X., Ren, S., et al., 2015, Delving deep into rectifiers: Surpassing human-level performance on imagenet classification: 2015 IEEE International Conference on Computer Vision, 1026–1034.
Hochreiter, S., and Schmidhuber, J., 1997, Long short-term memory: Neural Computation, 9(8), 1735–1780.
Hole, J. A., 1992, Nonlinear high-resolution three-dimensional seismic travel time tomography: Journal of Geophysical Research: Solid Earth, 97(B5), 6553–6562.
Huang, P., 2020, The research of damage identification method for bridge based on CNN-LSTM architecture neural network: MS Thesis, Qinghai University, Xining.
Ioffe, S., and Szegedy, C., 2015, Batch normalization: Accelerating deep network training by reducing internal covariate shift: Proceedings of the 32nd International Conference on Machine Learning, Lille, France, 37, 448–456.
Kazei, V., Ovcharenko, O., Plotnitskii, P., et al., 2021, Mapping full seismic waveforms to vertical velocity profiles by deep learning: Geophysics, 86(5), 1–50.
Kingma, D. P., and Ba, J., 2014, Adam: A method for stochastic optimization: arXiv: 1412.6980. [Online]. Available: https://arxiv.org/abs/1412.6980
LeCun, Y., Bengio, Y., and Hinton, G., 2015, Deep learning: Nature, 521(7553), 436–444.
LeCun, Y., Boser, B., Denker, J. S., et al., 1989, Backpropagation applied to handwritten zip code recognition: Neural Computation, 1(4), 541–551.
Lian, S., Yuan, S., Wang, G., et al., 2018, Enhancing low-wavenumber components of full-waveform inversion using an improved wavefield decomposition method in the time-space domain: Journal of Applied Geophysics, 157, 10–22.
Lin, Y., and Huang, L., 2014, Acoustic-and elastic-waveform inversion using a modified total-variation regularization scheme: Geophysical Journal International, 200(1), 489–502.
Liu, B., Yang, S., Ren, Y., et al., 2021, Deep-learning seismic full-waveform inversion for realistic structural models: Geophysics, 86(1), R31–R44.
Liu, Z., and Bleistein, N., 1995, Migration velocity analysis: Theory and an iterative algorithm: Geophysics, 60(1), 142–153.
Li, S., Liu, B., Ren, Y., et al., 2020, Deep-learning inversion of seismic data: IEEE Transactions on Geoscience and Remote Sensing, 58(3), 2135–2149.
Li, X., Aravkin, A. Y., van Leeuwen, T., et al., 2012, Fast randomized full-waveform inversion with compressive sensing: Geophysics, 77(3), A13–A17.
Maas, A. L., Hannun, A. Y., and Ng, A. Y., 2013, Rectifier nonlinearities improve neural network acoustic models: Proceedings of the 30th International Conference on Machine Learning, Atlanta, Georgia, USA, 30(1), 3.
Mao, B., Han, L. G., Feng, Q., et al, 2019, Subsurface velocity inversion from deep learning-based data assimilation: Journal of Applied Geophysics, 167, 172–179.
Meng, Z., and Scales, J. A., 1996, 2D tomography in multi-resolution analysis model space: 66th Annual International Meeting, SEG, Expanded Abstracts, 1126–1129.
Nair, V., and Hinton, G. E., 2010, Rectified linear units improve restricted Boltzmann machines: Proceedings of the 27th International Conference on Machine Learning, Haifa, Israel, 807–814.
Pratt, R. G., Shin, C., and Hick, G. J., 1998, Gauss-Newton and full Newton methods in frequency-space seismic waveform inversion: Geophysical Journal International, 133(2), 341–362.
Pytorch Official Document. [Online]. Available: https://pytorch.org/docs/stable/nn.init.html#
Saxe, A. M., McClelland, J. L., and Ganguli, S., 2013, Exact solutions to the nonlinear dynamics of learning in deep linear neural networks: arXiv: 1312.6120. [Online]. Available: https://arxiv.org/abs/1312.6120
Sun, J., Innanen, K. A., and Huang, C., 2021, Physics-guided deep learning for seismic inversion with hybrid training and uncertainty analysis: Geophysics, 86(3), R303–R317.
Sun, J., Niu, Z., Innanen, K. A., et al., 2020, A theory-guided deep-learning formulation and optimization of seismic waveform inversion: Geophysics, 85(2), R87–R99.
Woodward, M. J., 1992, Wave-equation tomography: Geophysics, 57(1), 15–26.
Acknowledgments
The authors are grateful to the reviewers for their valuable comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Cao Wei received a B.S. degree in information management and information system from Northeast Petroleum University, Daqing, China, in 2018. She is currently pursuing her Ph.D. degree at Northeast Petroleum University, Daqing, majoring in geological resources and geological engineering. Her research interest is seismic velocity inversion based on artificial intelligence. E-mail: caowei202007@163.com
Guo Xue-Bao (Corresponding author) received his B.S. and M.S. degrees in exploration technique and engineering from Northeast Petroleum University, Daqing, China, in 2012 and 2015, respectively, and a Ph.D. degree in solid geophysics from the Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China, in 2018. His research interests include reverse time migration and full waveform inversion. E-mail: guoxuebao1108@163.com
This research is financially supported by the Key Project of National Natural Science Foundation of China (No. 41930431), the Project of National Natural Science Foundation of China (Nos. 41904121, 41804133, and 41974116) and Joint Guidance Project of Natural Science Foundation of Heilongjiang Province (No. LH2020D006).
Rights and permissions
About this article
Cite this article
Wei, C., Xue-Bao, G., Feng, T. et al. Seismic velocity inversion based on CNN-LSTM fusion deep neural network. Appl. Geophys. 18, 499–514 (2021). https://doi.org/10.1007/s11770-021-0913-3
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11770-021-0913-3