Skip navigation

DSpace

機構典藏 DSpace 系統致力於保存各式數位資料(如:文字、圖片、PDF)並使其易於取用。

點此認識 DSpace
DSpace logo
English
中文
  • 瀏覽論文
    • 校院系所
    • 出版年
    • 作者
    • 標題
    • 關鍵字
  • 搜尋 TDR
  • 授權 Q&A
    • 我的頁面
    • 接受 E-mail 通知
    • 編輯個人資料
  1. NTU Theses and Dissertations Repository
  2. 電機資訊學院
  3. 資訊工程學系
請用此 Handle URI 來引用此文件: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70701
完整後設資料紀錄
DC 欄位值語言
dc.contributor.advisor林軒田(Hsuan-Tien Lin)
dc.contributor.authorBo-Yo Chenen
dc.contributor.author陳柏佑zh_TW
dc.date.accessioned2021-06-17T04:35:25Z-
dc.date.available2018-08-14
dc.date.copyright2018-08-14
dc.date.issued2018
dc.date.submitted2018-08-09
dc.identifier.citation[1] Cooperative Institute for Meteorological Satellite Studies. http://tropic.ssec.wisc. edu/tropic.php. [Online; accessed 8-February-2018].
[2] D. J. Cecil and E. J. Zipser. Relationships between tropical cyclone intensity and satellite-based indicators of inner core convection: 85-ghz ice-scattering signature and lightning. Monthly weather review, 127(1):103–123, 1999.
[3] V. F. Dvorak. Tropical cyclone intensity analysis and forecasting from satellite im- agery. Monthly Weather Review, 103(5):420–430, 1975.
[4] G. Fetanat, A. Homaifar, and K. R. Knapp. Objective tropical cyclone intensity es- timation using analogs of spatial features in satellite data. Weather and Forecasting, 28(6):1446–1459, 2013.
[5] A. K. Inamdar and K. R. Knapp. Intercomparison of independent calibration tech- niques applied to the visible channel of the isccp b1 data. Journal of Atmospheric and Oceanic Technology, 32(6):1225–1240, 2015.
[6] S.IoffeandC.Szegedy.Batchnormalization:Acceleratingdeepnetworktrainingby reducing internal covariate shift. In International conference on machine learning, pages 448–456, 2015.
[7] R. J. Joyce, J. E. Janowiak, P. A. Arkin, and P. Xie. Cmorph: A method that pro- duces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. Journal of Hydrometeorology, 5(3):487–503, 2004.
[8] S. Q. Kidder, M. D. Goldberg, R. M. Zehr, M. DeMaria, J. F. Purdom, C. S. Velden, N. C. Grody, and S. J. Kusselson. Satellite analysis of tropical cyclones using the advanced microwave sounding unit (amsu). Bulletin of the American Meteorological Society, 81(6):1241–1259, 2000.
[9] K.R.Knapp,S.Ansari,C.L.Bain,M.A.Bourassa,M.J.Dickinson,C.Funk,C.N. Helms, C. C. Hennon, C. D. Holmes, G. J. Huffman, et al. Globally gridded satellite observations for climate studies. Bulletin of the American Meteorological Society, 92(7):893–907, 2011.
[10] J. Kossin, K. Knapp, D. Vimont, R. Murnane, and B. Harper. A globally consistent reanalysis of hurricane variability and trends. Geophysical Research Letters, 34(4), 2007.
[11] A.KrizhevskyandG.Hinton.Learningmultiplelayersoffeaturesfromtinyimages. 2009.
[12] A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing sys- tems, pages 1097–1105, 2012.
[13] Y. LeCun. The mnist database of handwritten digits. http://yann. lecun. com/exdb/mnist/, 1998.
[14] X. Li, S. Chen, X. Hu, and J. Yang. Understanding the disharmony between dropout and batch normalization by variance shift. arXiv preprint arXiv:1801.05134, 2018.
[15] C.-C.Liu,C.-Y.Liu,T.-H.Lin,andL.-D.Chen.Asatellite-derivedtyphoonintensity index using a deviation angle technique. International Journal of Remote Sensing, 36(4):1216–1234, 2015.
[16] J. Miller, M. Maskey, and T. Berendes. Using deep learning for tropical cyclone intensity estimation. 2017.
[17] D. Mishkin and J. Matas. All you need is a good init. arXiv preprint arXiv:1511.06422, 2015.
[18] T. Nakazawa and S. Hoshino. Intercomparison of dvorak parameters in the tropical cyclone datasets over the western north pacific. Sola, 5:33–36, 2009.
[19] T. L. Olander and C. Velden. The current status of the uw-cimss advanced dvorak technique (adt). In 32nd Conf. on Hurricanes and Tropical Meteorology, 2012.
[20] T. L. Olander and C. S. Velden. The advanced dvorak technique: Continued devel- opment of an objective scheme to estimate tropical cyclone intensity using geosta- tionary infrared satellite imagery. Weather and Forecasting, 22(2):287–298, 2007.
[21] T. L. Olander and C. S. Velden. Tropical cyclone convection and intensity analy- sis using differenced infrared and water vapor imagery. Weather and Forecasting, 24(6):1558–1572, 2009.
[22] A. Perunicic. Choosing Weights: Small Changes, Big Differences. https://intoli. com/blog/neural-network-initialization/, 2017. [Online; accessed 25-July-2017].
[23] R. Pradhan, R. S. Aygun, M. Maskey, R. Ramachandran, and D. J. Cecil. Tropi- cal cyclone intensity estimation using a deep convolutional neural network. IEEE Transactions on Image Processing, 27(2):692–702, 2018.
[24] E. A. Ritchie, G. Valliere-Kelley, M. F. Piñeros, and J. S. Tyo. Tropical cyclone intensity estimation in the north atlantic basin using an improved deviation angle variance technique. Weather and Forecasting, 27(5):1264–1277, 2012.
[25] E. A. Ritchie, K. M. Wood, O. G. Rodríguez-Herrera, M. F. Piñeros, and J. S. Tyo. Satellite-derived tropical cyclone intensity in the north pacific ocean using the deviation-angle variance technique. Weather and Forecasting, 29(3):505–516, 2014.
[26] E. R. Sanabia, B. S. Barrett, and C. M. Fine. Relationships between tropical cyclone
intensity and eyewall structure as determined by radial profiles of inner-core infrared
brightness temperature. Monthly Weather Review, 142(12):4581–4599, 2014.
[27] K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale
image recognition. arXiv preprint arXiv:1409.1556, 2014.
[28] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke,
A. Rabinovich, et al. Going deeper with convolutions. Cvpr, 2015.
[29] C. Velden, B. Harper, F. Wells, J. L. Beven, R. Zehr, T. Olander, M. Mayfield, C. .
Guard, M. Lander, R. Edson, et al. The dvorak tropical cyclone intensity estimation
technique: A satellite-based method that has endured for over 30 years. Bulletin of
the American Meteorological Society, 87(9):1195–1210, 2006.
[30] C. Velden and D. Herndon. Update on the satellite consensus (satcon) algorithm for
estimating tc intensity. Poster session I. San Diego Google Scholar, 2014.
[31] C. S. Velden, T. L. Olander, and R. M. Zehr. Development of an objective scheme
to estimate tropical cyclone intensity from digital geostationary satellite infrared imagery.
Weather and Forecasting, 13(1):172–186, 1998.
[32] A. R. Zhai and J. H. Jiang. Dependence of us hurricane economic loss on maximum
wind speed and storm size. Environmental Research Letters, 9(6):064019, 2014.
[33] C.-J. Zhang, J.-F. Qian, L.-M. Ma, and X.-Q. Lu. Tropical cyclone intensity estimation
using rvm and dadi based on infrared brightness temperature. Weather and
Forecasting, 31(5):1643–1654, 2016.
[34] Y. Zhao, C. Zhao, R. Sun, and Z. Wang. A multiple linear regression model for
tropical cyclone intensity estimation from satellite infrared images. Atmosphere,
7(3):40, 2016.
dc.identifier.urihttp://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70701-
dc.description.abstract熱帶氣旋是發生於熱帶地區的極端天氣現象,強烈的熱帶氣旋所形成的颱風每年皆帶來可觀的經濟損失。對颱風強度的精準估計是防災的重要關鍵,更甚者,能讓颱風的生成與路徑預測更為準確。過去40年來,大氣研究員主要使用Dvorak法來估計颱風強度,然而此種方法在判讀衛星資料時得仰賴有經驗的研究員的人為判斷,因而受限於人為判斷帶來的潛在誤差。
近年來,颱風強度的估計吸引了許多資料科學家投身其中,各式各樣的機器學習技法被應用於此項任務。然而,至今為止並無對應的指標性數據集供科學家們用來比較各個方法之間的利與弊。這篇論文的其中一個目標便是整合現有來自衛星的觀測資料,成立一個全新的指標性數據集供未來各領域的科學家共同研究此一重要議題。
與此同時,近年來卷積神經網路技術已經發展得相當成熟,並在圖片辨識任務中取得豐碩的成果。我們以經典的卷積神經網路結構作為基礎,結合大氣科學家的專業知識,並根據颱風強度估計的任務特性做了些許特化,建立了一套全自動的颱風強度估計模型。相較於現行使用的系統,此應用了深度學習的模型表現更加穩定、全面,且同時降低人為判斷所帶來的潛在誤差,領先於全球,乃目前為止颱風強度相關研究在人工智慧上標誌性的重大突破。
此研究結果更充分顯示了深度學習在颱風科學上進一步發展的潛力,我們可將此深度學習模型應用在颱風強度預報、颱風形成預測等議題,應用範圍將不侷限於颱風強度的估計。其中颱風生成的預測更是現今大氣科學界尚未清楚了解的重要議題,我們或能以此研究作為基礎、仰賴人工智慧來探索,協助人類找出影響颱風形成的關鍵要素。
zh_TW
dc.description.abstractTropical cyclone (TC) is a type of severe weather systems that occur in tropical regions. Accurate estimation of TC intensity is crucial for disaster management. Moreover, the intensity estimation task is the key to understand and forecast the behavior of TCs better. For over 40 years, the Dvorak technique has been applied primarily for estimating TC intensity by meteorologists worldwide. However, these techniques are suffering from several issues, such as the inconsistency across various basins and the uncertainty raised by the inherent subjectivity.
Recently, the task has begun to attract attention from not only meteorologists but also data scientists. Nevertheless, it is hard to stimulate joint research between both types of scholars without a benchmark dataset to work on together. In this work, we release a such a benchmark dataset, which is a new open dataset collected from satellite remote sensing, for the TC-image-to-intensity estimation task.
We also propose a novel model to solve this task based on the convolutional neural network (CNN). We discover that the usual CNN, which is mature for object recognition, requires several modifications when being used for the intensity estimation task. Furthermore, we combine the domain knowledge of meteorologists, such as the rotation-invariance of TCs, into our model design to reach better performance. Experimental results on the released benchmark dataset verify that the proposed model not only outperforms the state-of-the-art operational techniques used nowadays for TC intensity estimation, but also is more stable across all situations. The results demonstrate the potential of applying data science for meteorology study.
en
dc.description.provenanceMade available in DSpace on 2021-06-17T04:35:25Z (GMT). No. of bitstreams: 1
ntu-107-R05922050-1.pdf: 1890853 bytes, checksum: 076fa49e24db7624bac6d146ee0fdcab (MD5)
Previous issue date: 2018
en
dc.description.tableofcontents誌謝iii
摘要v
Abstract vii
1 Introductions 1
2 Related work 3
3 The Tropical Cyclone for Image-to-intensity Regression dataset (TCIR) 7
3.1 Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.4 Other TC information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4 Proposed Method 13
4.1 Comparison to image classification . . . . . . . . . . . . . . . . . . . . . 13
4.1.1 Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1.2 Target . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1.3 Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1.4 Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3 Removal of Pooling and Dropout . . . . . . . . . . . . . . . . . . . . . . 17
4.3.1 Removal of Pooling . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.2 Removal of Dropout . . . . . . . . . . . . . . . . . . . . . . . . 19
4.4 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.5 Blending by Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.6 Regional Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5 Experiment and Analysis 23
5.1 Experiment Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.2 Dropout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.3 Rotate and Blending . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.4 Channel Combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.5 Smoothing the Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.6 Flipping The Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.7 Adding Regional Features . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.8 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6 Conclusion 37
Bibliography 39
dc.language.isoen
dc.title特化型卷積神經網路於颱風強度估計與預報的應用zh_TW
dc.titleTropical Cyclone Intensity Estimation With Specialized Convolutional Neural Networken
dc.typeThesis
dc.date.schoolyear106-2
dc.description.degree碩士
dc.contributor.oralexamcommittee林守德(Shou-De Lin),陳縕儂(Yun-Nung Chen)
dc.subject.keyword深度學習,大氣科學,颱風強度,卷積神經網路,zh_TW
dc.subject.keywordDeep learning,Atmospheric Science,Tropical cyclone Intensity,Convolutional Neural Network,Regression,Blending,Dropout,Pooling,en
dc.relation.page42
dc.identifier.doi10.6342/NTU201800988
dc.rights.note有償授權
dc.date.accepted2018-08-09
dc.contributor.author-college電機資訊學院zh_TW
dc.contributor.author-dept資訊工程學研究所zh_TW
顯示於系所單位:資訊工程學系

文件中的檔案:
檔案 大小格式 
ntu-107-1.pdf
  目前未授權公開取用
1.85 MBAdobe PDF
顯示文件簡單紀錄


系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。

社群連結
聯絡資訊
10617臺北市大安區羅斯福路四段1號
No.1 Sec.4, Roosevelt Rd., Taipei, Taiwan, R.O.C. 106
Tel: (02)33662353
Email: ntuetds@ntu.edu.tw
意見箱
相關連結
館藏目錄
國內圖書館整合查詢 MetaCat
臺大學術典藏 NTU Scholars
臺大圖書館數位典藏館
本站聲明
© NTU Library All Rights Reserved