利用深度學習辨識地震波 P 與 S 到時以及後續關聯與定位

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：136

、訪客IP：3.140.197.140

姓名

張立衡(Chang,Li-Heng) 查詢紙本館藏

畢業系所

地球科學學系

論文名稱

利用深度學習辨識地震波 P 與 S 到時以及後續關聯與定位
(Picking P and S phase based on deep learning and followed by associating and locating)

相關論文

★ 利用高解析反射與折射震測探測池上斷層

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2027-6-24以後開放)

摘要(中)

台灣位於環太平洋地震帶，每個月都會有數以千計的地震發生。近年來，儀器的廣
設與儲存資料的技術進步，使得成功記錄並且儲存到的地震數量大量提升。目前台灣每
年有超過18000個地震被記錄到。而人工智慧技術的迅速發展逐漸能夠取代例行性作業，
以減少人力需求並同時大量提高速度。
本研究主要利用 python 中的 obspy 套件做為主要的 I/O 套件，並且使用 keras 作為
深度學習的框架。在研究中，我們利用 Mousavi 等人在 2020 年提出的 Earthquake
transformer 模型作為基礎進行修改與優化，其訓練的基礎建立在為 P 波到時與 S 波到
時設計相對應的標籤。另外，本研究將原先的地震標籤改為雜訊標籤，並且在模型的輸
出層後加上了辨識器，使其成為一個生成對抗模型，與原先的 Earthquake transformer 相
比，有著更加穩定的輸出。在獲得地震到時後，本研究將 Chang 等人於 2019 發表的地
震事件與相關波形連結程式進行優化，加入本研究的資料處理流程中，完成了蒐集連續
資料到輸出定位結果的完整流程。
本研究中使用了和平 2017 年的資料進行訓練，並且使用和平 2020 年的資料做為測
試，並將整體流程應用在氣象局 2012 年一月的資料上。在和平 2020 年的資料中，P 波
與 S 波的到時召回率分別為 0.95 與 0.85，而精確度則分別為 0.87 與 0.81。透過本研究
開發的整體流程，能夠快速處理大量連續資料，並且在未來，能夠透過完整流程逐漸增
加訓練資料集，並在次投入訓練。使模型效能再次提升。待模型效能經多次檢驗後取得
一定的可靠度，便能投入既有的工作流程。

摘要(英)

Thousands of earthquakes occur in Taiwan because Taiwan locates in the circum-pacific
seismic belt. In recent years, with numerous settings of the seismic sensor and the improvement
of storage technology, the number of earthquakes that have been recorded and stored increased
a lot. An average of 18000 earthquakes are recorded each year. With the development of science
and technology, artificial intelligence can handle some single task to increase the processing
speed while reducing the manpower requirements.
In our research, we use the model structure which is modified and optimized based on the
model proposed from the study by Mousavi et al. in 2020. Training model based on designing
the label for P and S onset time. In addition, the original label for the seismic signal is changed
to the noise label, and also adds a discriminator after the final layer of the origin model. Make
the entire model structure become a generative adversarial network. Compared with the
earthquake transformer, the model in our study has more stable performance. After getting the
time P and S onset, we optimize the associator algorithm proposed by Chang et al. in 2019 and
add it to our process flow. And locate the earthquake by seisan then output the result in s-file
format.
We use the seismic waveform in Hoping during 2017 as training data, and use the data in
the same region during 2020 as testing data. We also applied the workflow to the continuous
data from CWB during January 2012. The recall of P and S onset in 2020 reaches 0.95 and 0.85,
and the precision reaches 0.87 and 0.81.
The workflow in our research can analyze a large amount of data very fast. In the future,
we also can increase the training dataset by the workflow. After the performance becomes stable,
we can add the flow to our other research, too.

關鍵字(中)

★ 深度學習
★ 關聯
★ 挑波

關鍵字(英)

論文目次

摘要................................................................................................................................... i
Abstract............................................................................................................................. ii
目錄................................................................................................................................. iii
圖目錄.............................................................................................................................. v
表目錄............................................................................................................................ vii
一、前言.................................................................................................................. 1
研究動機與目的...................................................................................... 1
文章架構.................................................................................................. 2
二、文獻回顧.......................................................................................................... 3
人工挑波基礎.......................................................................................... 3
傳統自動挑波.......................................................................................... 4
2-2-1 STA/LTA.............................................................................................. 4
2-2-2 AR-AIC............................................................................................... 4
深度學習方法.......................................................................................... 5
2-3-1 ARRU................................................................................................. 5
2-3-2 Earthquake transformer ................................................................... 5
三、研究方法.......................................................................................................... 7
類神經網路.............................................................................................. 7
3-1-1 神經元(Neuron)................................................................................ 7
3-1-2 激發函數 (activate function) .......................................................... 8
3-1-3 損失函數 (Loss function) ................................................................ 9
3-1-4 梯度下降(Gradient decent).............................................................. 9
3-1-5 反向傳遞 (Back propagation)........................................................ 10
3-1-6 優化器 (Optimizer)........................................................................ 10
卷積神經網路(Convolution Neural Network, CNN) .............................. 10
3-2-1 深度殘差網路 (Deep residual network, ResNet).......................... 11
遞迴神經網路(RNN) .............................................................................. 12
3-3-1 長短期記憶模型 (LSTM)............................................................... 12
3-3-2 自專注力機制 (Self attention mechanism) .................................. 13
生成對抗網路 (Generative Adversarial Network,GAN)........................ 13
評估方法................................................................................................ 13
3-5-1 混淆矩陣 (confusion matrix)......................................................... 13
3-5-2 精確率 (precision)、召回率 (recall)、F1 值............................... 14
關聯器(Associator)................................................................................. 15
3-6-1 到時聚合(Pick aggregation)........................................................... 15
3-6-2 生產候選事件(Event candidate creation)...................................... 15
iv
3-6-3 初始時間分析(Origin time analysis).............................................. 16
3-6-4 相位關聯(Phase association)......................................................... 16
3-6-5 資料輸出........................................................................................ 16
四、實驗流程........................................................................................................ 17
訓練目標................................................................................................ 17
資料........................................................................................................ 18
4-2-1 2017 年和平資料........................................................................... 19
4-2-2 2020 年和平資料........................................................................... 21
資料前處理............................................................................................ 23
模型設計................................................................................................ 23
模型訓練與測試.................................................................................... 26
4-5-1 已知事件測試................................................................................ 26
4-5-2 連續資料測試................................................................................ 26
資料評估................................................................................................ 26
五、實驗結果........................................................................................................ 27
2020 年和平資料 (已知事件)............................................................... 27
2020 年和平資料 (連續資料)............................................................... 32
氣象局資料 (連續資料)........................................................................ 32
關聯(associate)與定位結果................................................................... 33
5-4-1 2020 年和平資料 (連續資料) 定位結果..................................... 33
5-4-2 氣象局資料 (連續資料) 定位結果.............................................. 35
六、結果討論........................................................................................................ 36
標籤比較................................................................................................ 36
生成對抗模型(GAN)比較 ...................................................................... 40
加入雜訊訓練........................................................................................ 42
各模型比較............................................................................................ 46
訊噪比.................................................................................................... 49
和平 2020 年資料定位結果分析與比較.............................................. 49
6-6-1 和平 2020 年資料定位結果與分析.............................................. 50
氣象局到時結果分析............................................................................ 59
氣象局定位結果分析............................................................................ 62
七、結論................................................................................................................ 64
八、參考文獻........................................................................................................ 65

參考文獻

Chen Chen, Austin A. Holland; PhasePApy: A Robust PurePython Package for Automatic
Identification of Seismic Phases.Seismological Research Letters 2016; 87 (6): 1384-
1396. doi:https://doi.org/10.1785/0220160019
Chang, Y.-H., S.-H. Hung, and Y.-L. Chen, 2019: A fast algorithmfor automatic phase picker
and event location: Application to the 2018 Hualien earthquake sequences. Terr. Atmos.
Ocean. Sci., 30, 435-448, doi: 10.3319/TAO.2018.12.23.01
K. He, X. Zhang, S. Ren and J. Sun, "Deep Residual Learning for Image Recognition," 2016
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770-
778, doi: 10.1109/CVPR.2016.90.
Mitchell Withers, Richard Aster, Christopher Young, Judy Beiriger, Mark Harris, Susan Moore,
Julian Trujillo; A comparison of select trigger algorithms for automated global seismic
phase and event detection. Bulletin of the Seismological Society of America 1998; 88
(1): 95–106. doi: https://doi.org/10.1785/BSSA0880010095
Mousavi, S.M., Ellsworth, W.L., Zhu, W. et al. Earthquake transformer—an attentive deeplearning model for simultaneous earthquake detection and phase picking. Nat
Commun 11, 3952 (2020). doi: https://doi.org/10.1038/s41467-020-17591-w
Weiqiang Zhu, Gregory C Beroza, PhaseNet: a deep-neural-network-based seismic arrival-time
picking method, Geophysical Journal International, Volume 216, Issue 1, January 2019,
Pages 261–273, https://doi.org/10.1093/gji/ggy423
Wu‐Yu Liao, En‐Jui Lee, Dawei Mu, Po Chen, Ruey‐Juin Rau; ARRU Phase Picker:
Attention Recurrent‐Residual U‐Net for Picking Seismic P‐ and S‐Phase
Arrivals. Seismological Research Letters 2021;92(4):2410–2428.

指導教授

郭陳澔郭俊翔

審核日期

2022-6-27

推文