Yingtian Liu, Yong Li, Junheng Peng, Huating Li, Mingwei Wang
{"title":"时频相位混合域高分辨率闭环地震反演网络","authors":"Yingtian Liu, Yong Li, Junheng Peng, Huating Li, Mingwei Wang","doi":"arxiv-2408.04932","DOIUrl":null,"url":null,"abstract":"Thin layers and reservoirs may be concealed in areas of low seismic\nreflection amplitude, making them difficult to recognize. Deep learning (DL)\ntechniques provide new opportunities for accurate impedance prediction by\nestablishing a nonlinear mapping between seismic data and impedance. However,\nexisting methods primarily use time domain seismic data, which limits the\ncapture of frequency bands, thus leading to insufficient resolution of the\ninversion results. To address these problems, we introduce a new\ntime-frequency-phase (TFP) mixed-domain closed-loop seismic inversion network\n(TFP-CSIN) to improve the identification of thin layers and reservoirs. First,\nthe inversion network and closed-loop network are constructed by using\nbidirectional gated recurrent units (Bi-GRU) and convolutional neural network\n(CNN) architectures, enabling bidirectional mapping between seismic data and\nimpedance data. Next, to comprehensive learning across the entire frequency\nspectrum, the Fourier transform is used to capture frequency information and\nestablish frequency domain constraints. At the same time, the phase domain\nconstraint is introduced through Hilbert transformation, which improves the\nmethod's ability to recognize the weak reflection region features. Both\nexperiments on the synthetic data show that TFP-CSIN outperforms the\ntraditional supervised learning method and time domain semi-supervised learning\nmethods in seismic inversion. The field data further verify that the proposed\nmethod improves the identification ability of weak reflection areas and thin\nlayers.","PeriodicalId":501270,"journal":{"name":"arXiv - PHYS - Geophysics","volume":"4 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"High-resolution closed-loop seismic inversion network in time-frequency phase mixed domain\",\"authors\":\"Yingtian Liu, Yong Li, Junheng Peng, Huating Li, Mingwei Wang\",\"doi\":\"arxiv-2408.04932\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Thin layers and reservoirs may be concealed in areas of low seismic\\nreflection amplitude, making them difficult to recognize. Deep learning (DL)\\ntechniques provide new opportunities for accurate impedance prediction by\\nestablishing a nonlinear mapping between seismic data and impedance. However,\\nexisting methods primarily use time domain seismic data, which limits the\\ncapture of frequency bands, thus leading to insufficient resolution of the\\ninversion results. To address these problems, we introduce a new\\ntime-frequency-phase (TFP) mixed-domain closed-loop seismic inversion network\\n(TFP-CSIN) to improve the identification of thin layers and reservoirs. First,\\nthe inversion network and closed-loop network are constructed by using\\nbidirectional gated recurrent units (Bi-GRU) and convolutional neural network\\n(CNN) architectures, enabling bidirectional mapping between seismic data and\\nimpedance data. Next, to comprehensive learning across the entire frequency\\nspectrum, the Fourier transform is used to capture frequency information and\\nestablish frequency domain constraints. At the same time, the phase domain\\nconstraint is introduced through Hilbert transformation, which improves the\\nmethod's ability to recognize the weak reflection region features. Both\\nexperiments on the synthetic data show that TFP-CSIN outperforms the\\ntraditional supervised learning method and time domain semi-supervised learning\\nmethods in seismic inversion. The field data further verify that the proposed\\nmethod improves the identification ability of weak reflection areas and thin\\nlayers.\",\"PeriodicalId\":501270,\"journal\":{\"name\":\"arXiv - PHYS - Geophysics\",\"volume\":\"4 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Geophysics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.04932\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Geophysics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.04932","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
High-resolution closed-loop seismic inversion network in time-frequency phase mixed domain
Thin layers and reservoirs may be concealed in areas of low seismic
reflection amplitude, making them difficult to recognize. Deep learning (DL)
techniques provide new opportunities for accurate impedance prediction by
establishing a nonlinear mapping between seismic data and impedance. However,
existing methods primarily use time domain seismic data, which limits the
capture of frequency bands, thus leading to insufficient resolution of the
inversion results. To address these problems, we introduce a new
time-frequency-phase (TFP) mixed-domain closed-loop seismic inversion network
(TFP-CSIN) to improve the identification of thin layers and reservoirs. First,
the inversion network and closed-loop network are constructed by using
bidirectional gated recurrent units (Bi-GRU) and convolutional neural network
(CNN) architectures, enabling bidirectional mapping between seismic data and
impedance data. Next, to comprehensive learning across the entire frequency
spectrum, the Fourier transform is used to capture frequency information and
establish frequency domain constraints. At the same time, the phase domain
constraint is introduced through Hilbert transformation, which improves the
method's ability to recognize the weak reflection region features. Both
experiments on the synthetic data show that TFP-CSIN outperforms the
traditional supervised learning method and time domain semi-supervised learning
methods in seismic inversion. The field data further verify that the proposed
method improves the identification ability of weak reflection areas and thin
layers.