Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks

doi:10.1109/TGRS.2021.3066299

CORC > 海洋研究所 > 中国科学院海洋研究所 > 海洋环流与波动重点实验室

	Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks
	Wang, Chong 3; Zheng, Gang 2,8; Li, Xiaofeng 6,7; Xu, Qing 3,5; Liu, Bin 4; Zhang, Jun 1,9
刊名	IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
	2022
卷号	60 页码:16
关键词	Estimation Ocean temperature Clouds Tropical cyclones Cyclones Training Geostationary satellites Convolutional neural network (CNN) deep learning remote sensing tropical cyclone (TC)
ISSN号	0196-2892
DOI	10.1109/TGRS.2021.3066299
通讯作者	Li, Xiaofeng(xiaofeng.li@ieee.org) ; Xu, Qing(maggiexu@hhu.edu.cn)
英文摘要	In this study, a set of deep convolutional neural networks (CNNs) was designed for estimating the intensity of tropical cyclones (TCs) over the Northwest Pacific Ocean from the brightness temperature data observed by the Advanced Himawari Imager onboard the Himawari-8 geostationary satellite. We used 97 TC cases from 2015 to 2018 to train the CNN models. Several models with different inputs and parameters are designed. A comparative study showed that the selection of different infrared (IR) channels has a significant impact on the performance of the TC intensity estimate from the CNN models. Compared with the ground truth Best Track data of the maximum sustained wind speed, with a combination of four channels of data as input, the best multicategory CNN classification model has generated a fairly good accuracy (84.8x0025;) and low root mean square error (RMSE, 5.24 m/s) and mean bias (2.15 m/s) in TC intensity estimation. Adding attention layers after the input layer in the CNN helps to improve the model accuracy. The model is quite stable even with the influence of image noise. To reduce the side-effect of the very unbalanced distribution of TC category samples, we introduced a focalx005F;loss function into the CNN model. After we transformed the multiclassification problem into a binary classification problem, the accuracy increased to 88.9x0025;, and the RMSE and the mean bias are significantly reduced to 4.62 and x2212;0.76 m/s, respectively. The results show that our CNN models are robust in estimating TC intensity from geostationary satellite images.
资助项目	National Key Project of Research and Development Plan of China[2016YFC1401905] ; Chinese Academy of Sciences through the Strategic Priority Research Program[XDA19060101] ; Fundamental Research Funds for the Central Universities (Hohai University)[B200203026] ; Chinese Natural Science Foundation[41976163] ; Chinese Natural Science Foundation[41776183] ; Chinese Natural Science Foundation[41676167] ; Key Research and Development Project of Shandong Province[2019JZZY010102] ; Center for Ocean Mega-Science, CAS, through the Key Deployment Project[COMS2019R02] ; CAS Program[Y9KY04101L] ; Zhejiang Provincial Natural Science Foundation of China[LR21D060002]
WOS研究方向	Geochemistry & Geophysics ; Engineering ; Remote Sensing ; Imaging Science & Photographic Technology
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000730619400026
内容类型	期刊论文
源URL	[http://ir.qdio.ac.cn/handle/337002/177647]
专题	海洋研究所_海洋环流与波动重点实验室
通讯作者	Li, Xiaofeng; Xu, Qing
作者单位	1.Univ Miami, CIMAS, Miami, FL 33149 USA 2.Minist Nat Resources, State Key Lab Satellite Ocean Environm Dynam, Inst Oceanog 2, Hangzhou 310012, Peoples R China 3.Hohai Univ, Key Lab Marine Hazards Forecasting, Minist Nat Resources, Nanjing 210098, Peoples R China 4.Shanghai Ocean Univ, Sch Marine Sci, Shanghai 201306, Peoples R China 5.Ocean Univ China, Coll Informat Sci & Engn, Qingdao 266100, Peoples R China 6.Chinese Acad Sci, Ctr Ocean Megasci, Qingdao 266071, Peoples R China 7.Chinese Acad Sci, Inst Oceanog, CAS Key Lab Ocean Circulat & Waves, Qingdao 266071, Peoples R China 8.Southern Marine Sci & Engn Guangdong Lab Zhuhai, Zhuhai 519082, Peoples R China 9.Univ Miami, Hurricane Res Div, NOAA AOML, CIMAS, Miami, FL 33149 USA
推荐引用方式 GB/T 7714	Wang, Chong,Zheng, Gang,Li, Xiaofeng,et al. Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks[J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING,2022,60:16.
APA	Wang, Chong,Zheng, Gang,Li, Xiaofeng,Xu, Qing,Liu, Bin,&Zhang, Jun.(2022).Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks.IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING,60,16.
MLA	Wang, Chong,et al."Tropical Cyclone Intensity Estimation From Geostationary Satellite Imagery Using Deep Convolutional Neural Networks".IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 60(2022):16.