{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2023,10,7]],"date-time":"2023-10-07T17:14:34Z","timestamp":1696698874830},"reference-count":52,"publisher":"MDPI AG","issue":"19","license":[{"start":{"date-parts":[[2023,10,6]],"date-time":"2023-10-06T00:00:00Z","timestamp":1696550400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Stanford Earth Imaging Project"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"Building accurate acoustic subsurface velocity models is essential for successful industrial exploration projects. Traditional inversion methods from field-recorded seismograms struggle in regions with complex geology. While deep learning (DL) presents a promising alternative, its robustness using field data in these complicated regions has not been sufficiently explored. In this study, we present a thorough analysis of DL\u2019s capability to harness labeled seismograms, whether field-recorded or synthetically generated, for accurate velocity model recovery in a challenging region of the Gulf of Mexico. Our evaluation centers on the impact of training data selection and data augmentation techniques on the DL model\u2019s ability to recover velocity profiles. Models trained on field data produced superior results to data obtained using quantitative metrics like Mean Squared Error (MSE), Structural Similarity Index Measure (SSIM), and R2 (R-squared). They also yielded more geologically plausible predictions and sharper geophysical migration images. Conversely, models trained on synthetic data, while less precise, highlighted the potential utility of synthetic training data, especially when labeled field data are scarce. Our work shows that the efficacy of synthetic data-driven models largely depends on bridging the domain gap between training and test data through the use of advanced wave equation solvers and geologic priors. Our results underscore DL\u2019s potential to advance velocity model-building workflows in industrial settings using previously labeled field-recorded seismograms. They also highlight the indispensable role of earth scientists\u2019 domain expertise in curating synthetic data when field data are lacking.<\/jats:p>","DOI":"10.3390\/s23198277","type":"journal-article","created":{"date-parts":[[2023,10,6]],"date-time":"2023-10-06T12:29:12Z","timestamp":1696595352000},"page":"8277","source":"Crossref","is-referenced-by-count":0,"title":["Learning-Based Seismic Velocity Inversion with Synthetic and Field Data"],"prefix":"10.3390","volume":"23","author":[{"ORCID":"http:\/\/orcid.org\/0000-0003-1216-1963","authenticated-orcid":false,"given":"Stuart","family":"Farris","sequence":"first","affiliation":[{"name":"Department of Geophysics, Stanford University, Stanford, CA 94305, USA"}]},{"given":"Robert","family":"Clapp","sequence":"additional","affiliation":[{"name":"Department of Geophysics, Stanford University, Stanford, CA 94305, USA"}]},{"given":"Mauricio","family":"Araya-Polo","sequence":"additional","affiliation":[{"name":"TotalEnergies, EP R&T, Houston, TX 77002, USA"}]}],"member":"1968","published-online":{"date-parts":[[2023,10,6]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Yilmaz, \u00d6. (2001). Seismic Data Analysis: Processing, Inversion, and Interpretation of Seismic Data, Society of Exploration Geophysicists.","DOI":"10.1190\/1.9781560801580"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Sheriff, R.E., and Geldart, L.P. (1995). Exploration Seismology, Cambridge University Press.","DOI":"10.1017\/CBO9781139168359"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"109","DOI":"10.1016\/j.jappgeo.2010.05.004","article-title":"Monitoring CO2 response on surface seismic data; a rock physics and seismic modeling feasibility study at the CO2 sequestration site, Ketzin, Germany","volume":"71","author":"Kazemeini","year":"2010","journal-title":"J. Appl. Geophys."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"859","DOI":"10.1190\/1.1440058","article-title":"Velocity Spectra\u2013Digital Computer Derivation Applications of Velocity Functions","volume":"34","author":"Taner","year":"1969","journal-title":"Geophysics"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"903","DOI":"10.1190\/1.1441970","article-title":"Tomographic determination of velocity and depth in laterally varying media","volume":"50","author":"Bishop","year":"1985","journal-title":"Geophysics"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"483","DOI":"10.1190\/1.1443065","article-title":"Linear aspects of tomographic velocity analysis","volume":"56","author":"Stork","year":"1991","journal-title":"Geophysics"},{"key":"ref_7","unstructured":"Lailly, P. (1983, January 16\u201318). The seismic inverse problem as a sequence of before stack migrations. Proceedings of the Conference on Inverse Scattering, Tulsa, OK, USA."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"1259","DOI":"10.1190\/1.1441754","article-title":"Inversion of seismic reflection data in the acoustic approximation","volume":"49","author":"Tarantola","year":"1984","journal-title":"Geophysics"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"WCC1","DOI":"10.1190\/1.3238367","article-title":"An overview of full-waveform inversion in exploration geophysics","volume":"74","author":"Virieux","year":"2009","journal-title":"Geophysics"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"ref_11","unstructured":"Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep Learning, MIT Press. Available online: http:\/\/www.deeplearningbook.org."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1109\/MSP.2017.2760358","article-title":"Using deep neural networks for inverse problems in imaging: Beyond analytical methods","volume":"35","author":"Lucas","year":"2018","journal-title":"IEEE Signal Process. Mag."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"39","DOI":"10.1109\/JSAIT.2020.2991563","article-title":"Deep learning techniques for inverse problems in imaging","volume":"1","author":"Ongie","year":"2020","journal-title":"IEEE J. Sel. Areas Inf. Theory"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"128","DOI":"10.1109\/MSP.2019.2950640","article-title":"Deep-learning methods for parallel magnetic resonance imaging reconstruction: A survey of the current approaches, trends, and issues","volume":"37","author":"Knoll","year":"2020","journal-title":"IEEE Signal Process. Mag."},{"key":"ref_15","unstructured":"(2023, March 16). FDA Clears GE\u2019s Deep Learning Image Reconstruction Engine. Available online: https:\/\/www.itnonline.com\/content\/fdaclears-ges-deep-learning-image-reconstruction-engine."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Chen, C., Chen, Q., Xu, J., and Koltun, V. (2018, January 18\u201323). Learning to see in the dark. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00347"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Godard, C., Mac Aodha, O., and Brostow, G.J. (2017, January 21\u201326). Unsupervised monocular depth estimation with left-right consistency. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.","DOI":"10.1109\/CVPR.2017.699"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"1524","DOI":"10.1093\/gji\/ggx380","article-title":"Efficient monte carlo sampling of inverse problems using a neural network-based forward\u2014Applied to gpr crosshole traveltime inversion","volume":"211","author":"Hansen","year":"2017","journal-title":"Geophys. J. Int."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"53","DOI":"10.1007\/s11004-019-09832-6","article-title":"Stochastic seismic waveform inversion using generative adversarial networks as a geological prior","volume":"52","author":"Mosser","year":"2020","journal-title":"Math. Geosci."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"526","DOI":"10.1190\/tle38070526.1","article-title":"Applications of supervised deep learning for seismic interpretation and inversion","volume":"38","author":"Zheng","year":"2019","journal-title":"Lead. Edge"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Le, T.A., Baydin, A.G., Zinkov, R., and Wood, F.D. (2017, January 14\u201319). Using synthetic data to train neural networks is model-based reasoning. Proceedings of the IEEE International Joint Conference on Neural Network, Anchorage, AK, USA.","DOI":"10.1109\/IJCNN.2017.7966298"},{"key":"ref_22","unstructured":"Jaderberg, M., Simonyan, K., Vedaldi, A., and Zisserman, A. (2014). Synthetic data and artificial neural networks for natural scene text recognition. arXiv."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Moate, C.P., Hayward, S.D., Ellis, J.S., Russell, L., Timmerman, R.O., Lane, R.O., and Strain, T.J. (2018, January 27\u201329). Vehicle detection in infrared imagery using neural networks with synthetic training data. Proceedings of the Image Analysis and Recognition: 15th International Conference, ICIAR 2018, P\u00f3voa de Varzim, Portugal.","DOI":"10.1007\/978-3-319-93000-8_51"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"58","DOI":"10.1190\/tle37010058.1","article-title":"Deep-learning tomography","volume":"37","author":"Jennings","year":"2018","journal-title":"Lead. Edge"},{"key":"ref_25","unstructured":"Araya-Polo, M., Adler, A., Farris, S., and Jennings, J. (2019). Deep Learning: Architectures, Algorithms, and Applications, Springer."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Farris, S., Araya-Polo, M., Jennings, J., Clapp, B., and Biondi, B. (2018, January 21\u201322). Tomography: A deep learning vs full-waveform inversion comparison. Proceedings of the EAGE Workshop on High Performance Computing, Santander, Columbia.","DOI":"10.3997\/2214-4609.201803073"},{"key":"ref_27","unstructured":"Simon, J., Fabien-Ouellet, G., Gloaguen, E., Khurjekar, I., and Araya-Polo, M. (October, January 26). Velocity model building using transfer learning. Proceedings of the SEG\/AAPG\/SEPM First International Meeting for Applied Geoscience & Energy, Denver, CO, USA."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"89","DOI":"10.1109\/MSP.2020.3037429","article-title":"Deep learning for seismic inverse problems: Toward the acceleration of geophysical analysis workflows","volume":"38","author":"Adler","year":"2021","journal-title":"IEEE Signal Process. Mag."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Gelboim, M., Adler, A., Sun, Y., and Araya-Polo, M. (2022). Encoder\u2013Decoder Architecture for 3D Seismic Inversion. Sensors, 23.","DOI":"10.3390\/s23010061"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"R711","DOI":"10.1190\/geo2019-0473.1","article-title":"Mapping full seismic waveforms to vertical velocity profiles by deep learning","volume":"86","author":"Kazei","year":"2021","journal-title":"Geophysics"},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"V33","DOI":"10.1190\/geo2018-0870.1","article-title":"Automatic velocity analysis using convolutional neural network and transfer learning","volume":"85","author":"Park","year":"2020","journal-title":"Geophysics"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"U21","DOI":"10.1190\/geo2018-0786.1","article-title":"Seismic velocity estimation: A deep recurrent neural-network approach","volume":"85","author":"Sarkar","year":"2020","journal-title":"Geophysics"},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/TGRS.2022.3185794","article-title":"Multi-task learning for low-frequency extrapolation and elastic model building from seismic data","volume":"60","author":"Ovcharenko","year":"2022","journal-title":"IEEE Trans. Geosci. Remote. Sens."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"WA137","DOI":"10.1190\/geo2019-0340.1","article-title":"High-resolution reservoir characterization using deep learning-aided elastic full-waveform inversion: The North Sea field data example","volume":"85","author":"Zhang","year":"2020","journal-title":"Geophysics"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Arts, R., Chadwick, A., Eiken, O., Thibeau, S., and Nooner, S. (2008). Ten years\u2019 experience of monitoring CO2 injection in the Utsira Sand at Sleipner, offshore Norway. First Break, 26.","DOI":"10.3997\/1365-2397.26.1115.27807"},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Ringrose, P. (2020). How to Store CO2 Underground: Insights from Early-Mover CCS Projects, Springer.","DOI":"10.1007\/978-3-030-33113-9"},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"315","DOI":"10.1144\/petgeo2016-050","article-title":"Hydrogen storage in a heterogeneous sandstone formation: Dimensioning and induced hydraulic effects","volume":"23","author":"Pfeiffer","year":"2017","journal-title":"Pet. Geosci."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"23436","DOI":"10.1016\/j.ijhydene.2020.08.138","article-title":"Underground hydrogen storage: A comprehensive review","volume":"46","author":"Zivar","year":"2021","journal-title":"Int. J. Hydrog. Energy"},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"2666","DOI":"10.1190\/1.1441889","article-title":"State-of-the-art geophysical exploration for geothermal resources","volume":"50","author":"Wright","year":"1985","journal-title":"Geophysics"},{"key":"ref_40","doi-asserted-by":"crossref","first-page":"127","DOI":"10.1016\/S0040-1951(02)00668-6","article-title":"Extensional shear zones as imaged by reflection seismic lines: The Larderello geothermal field (central Italy)","volume":"363","author":"Brogi","year":"2003","journal-title":"Tectonophysics"},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"35","DOI":"10.1016\/j.geothermics.2009.12.004","article-title":"Exploring the Gro\u00df Sch\u00f6nebeck (Germany) geothermal site using a statistical joint interpretation of magnetotelluric and seismic tomography models","volume":"39","author":"Munoz","year":"2010","journal-title":"Geothermics"},{"key":"ref_42","first-page":"101","article-title":"MLReal: Bridging the gap between training on synthetic data and real data applications in machine learning","volume":"3","author":"Alkhalifah","year":"2022","journal-title":"Artif. Intell. Geosci."},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Naranjo, J.C., Ekstrand, E.J., Etgen, J., Hargrove, K.L., Brewton, J.G., Garcia, O., Astvatsaturov, G., Hite, D., Howieson, B., and Bai, B. (2011, January 18\u201323). Survey design and implementation of a multiple wide-azimuth towed streamer seismic acquisition strategy at the tiber discovery; deep water Gulf of Mexico, USA. Proceedings of the Society of Exploration Geophysicists International Exposition and 81st Annual Meeting 2011, SEG 2011, San Antonio, TX, USA.","DOI":"10.1190\/1.3627449"},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"\u00c7i\u00e7ek, \u00d6., Abdulkadir, A., Lienkamp, S.S., Brox, T., and Ronneberger, O. (2016, January 17\u201321). 3D U-Net: Learning dense volumetric segmentation from sparse annotation. Proceedings of the Medical Image Computing and Computer-Assisted Intervention\u2013MICCAI 2016: 19th International Conference, Athens, Greece.","DOI":"10.1007\/978-3-319-46723-8_49"},{"key":"ref_45","unstructured":"Aki, K., and Richards, P.G. (2002). Quantitative Seismology, University Science Books. [2nd ed.]."},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"770","DOI":"10.1190\/1.1440465","article-title":"Formation velocity and density\u2014the diagnostic basis for stratigraphic traps","volume":"39","author":"Gardner","year":"1974","journal-title":"Geophysics"},{"key":"ref_47","unstructured":"Clapp, R.G. (2018). Synthetic Model Building for Training Neural Networks in a Jupyter Notebook, SEP. Technical report."},{"key":"ref_48","doi-asserted-by":"crossref","first-page":"R235","DOI":"10.1190\/geo2012-0382.1","article-title":"A critical appraisal of asymptotic 3D-to-2D data transformation in full-waveform seismic crosshole tomography","volume":"78","author":"Auer","year":"2013","journal-title":"Geophysics"},{"key":"ref_49","first-page":"261","article-title":"Structure-oriented smoothing and semblance","volume":"635","author":"Hale","year":"2009","journal-title":"CWP Rep."},{"key":"ref_50","doi-asserted-by":"crossref","first-page":"600","DOI":"10.1109\/TIP.2003.819861","article-title":"Image quality assessment: From error visibility to structural similarity","volume":"13","author":"Wang","year":"2004","journal-title":"IEEE Trans. Image Process."},{"key":"ref_51","doi-asserted-by":"crossref","unstructured":"Farris, S., Barnier, G., and Clapp, R. (2022, January 25\u201330). Deep learning velocity model building using an ensemble regression approach. Proceedings of the SEG\/AAPG International Meeting for Applied Geoscience & Energy, Houston, TX, USA.","DOI":"10.1190\/image2022-w7-01.1"},{"key":"ref_52","doi-asserted-by":"crossref","first-page":"1514","DOI":"10.1190\/1.1441434","article-title":"Reverse time migration","volume":"48","author":"Baysal","year":"1983","journal-title":"Geophysics"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/23\/19\/8277\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,10,6]],"date-time":"2023-10-06T12:34:30Z","timestamp":1696595670000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/23\/19\/8277"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,10,6]]},"references-count":52,"journal-issue":{"issue":"19","published-online":{"date-parts":[[2023,10]]}},"alternative-id":["s23198277"],"URL":"https:\/\/doi.org\/10.3390\/s23198277","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,10,6]]}}}