A Meta-Learning-Based Approach for Automatic First-Arrival Picking

被引:0
作者
Li, Hanyang [1 ,2 ]
Sun, Yuhang [2 ]
Li, Jiahui [2 ]
Li, Hang [2 ]
Dong, Hongli [3 ,4 ]
机构
[1] Northeast Petr Univ, Sanya Offshore Oil & Gas Res Inst, Sanya 572025, Peoples R China
[2] Northeast Petr Univ, Artificial Intelligence Energy Res Inst, Daqing 163318, Peoples R China
[3] Northeast Petr Univ, Artificial Intelligence Energy Res Inst, Natl Key Lab Continental Shale Oil, Daqing 163318, Peoples R China
[4] Northeast Petr Univ, Heilongjiang Prov Key Lab Networking & Intelligent, Daqing 163318, Peoples R China
来源
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING | 2024年 / 62卷
基金
中国国家自然科学基金;
关键词
Training; Task analysis; Metalearning; Data models; Geology; Adaptation models; Accuracy; First-arrival picking; meta-learning; nonaccurate label; seismic data; PHASE ARRIVAL TIMES; CROSS-CORRELATION;
D O I
10.1109/TGRS.2024.3436817
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
Precise first-arrival picking holds pivotal importance in the realms of seismic data processing and microseismic monitoring. Recently, data-driven approaches have shown remarkable performance. However, these approaches rely on high-quality labeled datasets and involve a time-consuming and labor-intensive labeling process. In addition, data-driven picking methods often suffer from generalization problems in the face of varying noise characteristics and geological environments. To tackle the challenges head-on, this study introduces a novel training algorithm grounded in meta-learning. In contrast to traditional training methods, this innovative approach distinguishes itself by reducing the costs associated with dataset creation and requiring only a modest number of high-quality labeled samples to achieve superior performance. Furthermore, the proposed method can be seamlessly implemented with different types of deep neural networks (DNNs). Our extensive experimentation on two field datasets encompassing distinct geological zones demonstrates the method's effectiveness in alleviating the dependence on high-quality training samples, enhancing first-arrival picking accuracy, and bolstering the model's robustness against strong noise interference.
引用
收藏
页数:15
相关论文
共 44 条
[1]   Refinement of arrival-time picks using a cross-correlation based workflow [J].
Akram, Jubran ;
Eaton, David W. .
JOURNAL OF APPLIED GEOPHYSICS, 2016, 135 :55-66
[2]  
Al-Shedivat Maruan, 2021, P 24 INT C ARTIFICIA, V130, P1369
[3]  
ALLEN RV, 1978, B SEISMOL SOC AM, V68, P1521
[4]  
Behl H.S., 2019, arXiv
[5]   Multiple Meta-model Quantifying for Medical Visual Question Answering [J].
Do, Thong ;
Nguyen, Binh X. ;
Tjiputra, Erman ;
Tran, Minh ;
Tran, Quang D. ;
Anh Nguyen .
MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2021, PT V, 2021, 12905 :64-74
[6]   Meta-Learning of Neural Architectures for Few-Shot Learning [J].
Elsken, Thomas ;
Staffler, Benedikt ;
Metzen, Jan Hendrik ;
Hutter, Frank .
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2020), 2020, :12362-12372
[7]   AEnet: Automatic Picking of P-Wave First Arrivals Using Deep Learning [J].
Guo, Chao ;
Zhu, Tieyuan ;
Gao, Yongtao ;
Wu, Shunchuan ;
Sun, Jian .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2021, 59 (06) :5293-5303
[8]   First-Arrival Picking for Microseismic Monitoring Based on Deep Learning [J].
Guo, Xiaolong .
INTERNATIONAL JOURNAL OF GEOPHYSICS, 2021, 2021
[9]   Meta-Learning in Neural Networks: A Survey [J].
Hospedales, Timothy ;
Antoniou, Antreas ;
Micaelli, Paul ;
Storkey, Amos .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2022, 44 (09) :5149-5169
[10]   Automated determination of P-phase arrival times at regional and local distances using higher order statistics [J].
Kueperkoch, L. ;
Meier, T. ;
Lee, J. ;
Friederich, W. .
GEOPHYSICAL JOURNAL INTERNATIONAL, 2010, 181 (02) :1159-1170