Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | 변중무 | - |
dc.date.accessioned | 2022-07-27T04:50:54Z | - |
dc.date.available | 2022-07-27T04:50:54Z | - |
dc.date.issued | 2020-10 | - |
dc.identifier.citation | SEG Technical Program Expanded Abstracts 2020, page. 616-620 | en_US |
dc.identifier.issn | 1949-4645 | - |
dc.identifier.issn | 1052-3812 | - |
dc.identifier.uri | https://library.seg.org/doi/10.1190/segam2020-3427240.1 | - |
dc.identifier.uri | https://repository.hanyang.ac.kr/handle/20.500.11754/171821 | - |
dc.description.abstract | The conventional interpretation of airborne electromagnetic (AEM) data has been conducted by solving the inverse problem. With the recent advance in machine learning (ML) techniques, a one-dimensional (1D) deep neural network (DNN) inversion scheme which predicts a 1D resistivity model using multi-frequency vertical magnetic fields and altitude information at one location was suggested. The final image of this 1D approach was constructed by connecting 1D resistivity models. However, 1D ML interpretation shows the low performance in accurate estimation of a conductive anomaly like 1D conventional inversion. Thus, we suggest a two-dimensional (2D) interpretation technique, which can consider spatial continuity by using the recurrent neural network (RNN). We generated various 2D resistivity models and calculated vertical magnetic fields, then trained the RNN by corresponding EM responses and resistivity models. To verify the RNN inversion scheme, we applied to the trained RNN to the synthetic and field data. The inversion result of field data matched well with the conventional inversion results. In addition, compared to the 1D DNN, RNN inversion showed better resolution for an isolated conductive anomaly. | en_US |
dc.description.sponsorship | This work was supported by a Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant, funded by the Korean government (MOTIE) (20194010201920). This work was also supported by the Reservoir Imaging with Seismic & EM technology using Machine Learning (RISE.ML) Consortium at Hanyang University. Also, we wish to thank UBC-GIF for sharing the DIGHEM dataset. | en_US |
dc.language.iso | en | en_US |
dc.publisher | Society of Exploration Geophysicist | en_US |
dc.subject | Geology | en_US |
dc.subject | Artificial intelligence | en_US |
dc.subject | business.industry | en_US |
dc.subject | business | en_US |
dc.subject | Recurrent neural network | en_US |
dc.subject | Pattern recognition | en_US |
dc.title | Imaging of subsurface orebody with airborne electromagnetic data using a recurrent neural network | en_US |
dc.type | Article | en_US |
dc.identifier.doi | 10.1190/segam2020-3427240.1 | - |
dc.relation.page | 616-620 | - |
dc.contributor.googleauthor | Bang, Minkyu | - |
dc.contributor.googleauthor | Oh, Seokmin | - |
dc.contributor.googleauthor | Noh, Kyubo | - |
dc.contributor.googleauthor | Seol, Soon Jee | - |
dc.contributor.googleauthor | Byun, Joongmoo | - |
dc.sector.campus | S | - |
dc.sector.daehak | COLLEGE OF ENGINEERING[S] | - |
dc.sector.department | DEPARTMENT OF EARTH RESOURCES AND ENVIRONMENTAL ENGINEERING | - |
dc.identifier.pid | jbyun | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.