[1]蔡育埼,邵 博,于子叶,等.基于PRIME-DP预训练模型的滑坡微地震等非天然地震分类研究[J].华南地震,2025,(01):12-18.[doi:10.13512/j.hndz.2025.01.02]
　CAI Yuqi,SHAO Bo,YU Ziye,et al.Classification of Non-Natural Seismic Events with Landslide Microseismicity as Representative Based on PRIME-DP Pre-Training Model[J].,2025,(01):12-18.[doi:10.13512/j.hndz.2025.01.02]
点击复制

基于PRIME-DP预训练模型的滑坡微地震等非天然地震分类研究()
分享到：

华南地震[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年01期
页码:
12-18
栏目:
地震科学研究
出版日期:
2025-03-30

文章信息/Info

Title:
Classification of Non-Natural Seismic Events with Landslide Microseismicity as Representative Based on PRIME-DP Pre-Training Model
文章编号:
1001-8662（2025）01-0012-07
作者:
蔡育埼1; 2; 邵 博1; 于子叶2; 姚翔龙1; 刘 路1; 欧阳金恵1
1.中国长江三峡集团科学技术研究院，武汉 430010;2.中国地震局地球物理研究所，北京 100081
Author(s):
CAI Yuqi1; 2; SHAO Bo1; YU Ziye2; YAO Xianglong1; LIU Lu1; OUYANG Jinhui1
1.Institute of Science and Technology ， China Three Gorges Corporation ， Wuhan 100089， China;2.Institute of Geophysics ， China Earthquake Administration ， Beijing 100081， China
关键词:
地震信号分类; 地震预训练模型; 地震大模型
Keywords:
Classification of seismic signal; Pre-training seismic model; Seismic large model
分类号:
P315.9
DOI:
10.13512/j.hndz.2025.01.02
文献标志码:
A
摘要:
天然地震和非天然事件分类对于地震编目、地震预警、滑坡灾害分析等具有重要意义。目前的研究方法中传统方法需要构建P/S振幅比、谱比等参数，这限制了算法对于复杂类别分类的可能。而基于深度神经网络的算法虽然可以对复杂的事件进行分类，但是目前研究中多集中于爆破、塌陷事件的识别，对于滑坡微地震事件等更多类型分类则研究较少。更重要的是深度神经网络训练需要大量人工标注数据，而非天然地震事件标注数据较为缺乏，因此难以训练高精度、高泛化能力的模型。因此我们基于开源的PRIME-DP地震数据处理预训练模型训练了用于天然地震、爆破、塌陷、滑坡微地震事件的多分类模型。并使用预训练模型数据和262条微地震事件进行了迁移学习训练。训练结果表明，我们的模型相比于基于STFT特征的深度学习分类模型精度从79.4%提升到了94.8%。这说明基于预训练模型来进行地震分类可以有效的提升最终精度。
Abstract:
Classification of natural earthquakes and non-natural seismic events is of great significance for earthquake cataloguing， earthquake early warning， and landslide disaster analysis. The existing traditional research methods need to construct parameters such as the P/S amplitude ratio and spectrum ratio，which limits the possibility of the algorithm to classify complex categories. Although the algorithms based on deep neural networks can classify complex events， most of the current research focuses on the identification of blasting and collapse events， and there is less research on the classification of more types such as landslide microseismicity. More importantly， deep neural network training requires a large amount of manually labeled data， which is relatively scarce for non-natural seismic event， so it is difficult to train models with high precision and high generalization capabilities. Therefore，this paper trained multi-classification models for natural earthquakes，blasting，collapse， and landslide microseismicity events based on the open-source PRIME-DP seismic data processing pre-training model. The transfer learning and training were performed by using the pre-training model data and 262 microseismicity events. The training results show that compared with the deep learning classification model based on STFT features， the accuracy of the proposed model is improved from 79.4% to 94.8%. This indicates that seismic event classification based on the pre-training model can effectively improve the final accuracy.

参考文献/References:

[1]蔡育埼，于子叶.基于迁移学习的岩石边坡微地震事件检测算法[J].地震地磁观测与研究，2024，45（2）：20-27.
[2] Yu Z，Wang W. LPPN：A lightweight network for fast phase picking[J]. Seismological Society of America，2022，93（5）：2834-2846.
[3] Yu Z，Wang W，Chen Y. Benchmark on the accuracy and efficiency of several neural network based phase pickers using datasets from China Seismic Network[J]. Earthquake Science，2023，36（2）：113-131.
[4] Scarpetta S，Giudicepietro F，Ezin E C，et al. Automatic classification of seismic signals at Mt. Vesuvius volcano， Italy，using neural networks[J]. Bulletin of the Seismological Society of America，2005，95（1）：185-196.
[5]边银菊.遗传BP网络在地震和爆破识别中的应用[J].地震学报，2002，24（5）：516-524+558.
[6] Kim W Y，Simpson D W，Richards P G. High-frequency spec?tra of regional phases from earthquakes and chemical explo?sions[J]. Bulletin of the Seismological Society of America， 1994，84（5）：1365-1386.
[7]任涛，林梦楠，陈宏峰，等.基于Bagging集成学习算法的地震事件性质识别分类[J].地球物理学报，2019，62（1）：383-392.
[8]王婷婷，边银菊，任梦依，等.地震事件分类识别软件[J].地震，2024，44（2）：104-119.
[9]杨千里，王婷婷，边银菊.基于广义S变换的地震与爆炸识别[J].地震学报，2020，42（5）：613-628.
[10] Esposito A M，Giudicepietro F，Scarpetta S，et al. Automatic discrimination among landslide，explosion-quake，and mi?crotremor seismic signals at Stromboli volcano using neural networks[J]. Bulletin of the Seismological Society of Ameri?ca，2006，96（4A）：1230-1240.
[11] Tang L，Zhang M，Wen L. Support vector machine classifica?tion of seismic events in the Tianshan orogenic belt[J]. Jour?nal of Geophysical Research：Solid Earth，2020，125（1）：e2019JB018132.
[12]陈润航，黄汉明，柴慧敏.地震和爆破事件源波形信号的卷积神经网络分类研究[J].地球物理学进展，2018，33 （4）：1331-1338.
[13]黎炳君，黄汉明，王婷婷，等.基于STFT和CNN的地震信号分类识别研究[J].地球物理学进展，2021，36（4）：1404-1411.
[14] Titos M，Bueno A，García L，et al. Detection and classifica?tion of continuous volcano-seismic signals with recurrent neural networks[J]. IEEE Transactions on Geoscience and Remote Sensing，2018，57（4）：1936-1948.
[15] Devlin J，Change M W，Lee K，et al. Bert：Pre-training of deep bidirectional transformers for language understanding [C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics：Human Language Technologies. Minneapolis， Minnesota：Association for Computational Linguistics，2019.
[16] Radford A，Narasimhan K，Salimans M，et al. Improving lan?guage understanding by generative pre-training[J]. OpenAI， 2018：1-12.
[17] Radford A，Wu J，Child R，et al. Language models are unsupervised multitask learners[J]. OpenAI blog，2019，1 （8）：9.
[18] Brown T，Mann B，Ryder N，et al. Language models are few-shot learners[J]. Advances in neural information processing systems，2020（33）：1877-1901.
[19] Yu Z，Cai Y，Wang W，et al. PRIME-DP：Pre-trained integrated model for earthquake data processing[J]. arXiv， 2024（08）：01919.

备注/Memo

备注/Memo:
收稿日期：2024-12-31
基金项目：中国长江三峡集团科学技术研究院相关基金（WWKY-2021-0273，202203016）
作者简介：蔡育埼（1999-），男，硕士，主要从事地震学机器学习算法相关研究。E-mail：caiyuqiming@foxmail.com
更新日期/Last Update: 2025-03-30