融合空间物理参数的小样本极光涡旋自动分类研究

韩冰; 吴墨豪; 胡泽骏; 王平; 张义生

doi:10.6038/cjg2023Q0657

融合空间物理参数的小样本极光涡旋自动分类研究

- 1.
  西安电子科技大学电子工程学院, 西安 710071
- 2.
  中国极地研究中心自然资源部极地科学重点实验室, 上海 200136
- 3.
  浙江大学海洋研究院, 浙江舟山 316021
- 4.
  北京应用气象研究所, 北京 100029
基金项目:
国家自然科学基金(62076190, 41831072, 41874195, 42074199), 国家重点研发计划(2021YFE0106400), 陕西省重点研发计划项目(2022ZDLGY01-11), 上海市科委项目(21DZ1206100), 中国科学院空间科学先导专项(XDA15350202), 空间环境地基综合监测网项目, 应用气象研究所基础研究项目

详细信息

作者简介:
韩冰, 女, 1978年生, 教授, 主要从事信号处理、视觉感知、计算机视觉、模式识别等方向研究.E-mail: bhan@xidian.edu.cn

通讯作者: 胡泽骏, 男, 1978年生, 研究员, 博士生导师, 研究方向为极光物理, 极区太阳风-磁层-电离层耦合, 极区空间环境/空间天气方面的研究.E-mail: huzejun@pric.org.cn

中图分类号: P353

收稿日期: 2022-08-17

修回日期: 2023-04-11

上线日期: 2023-11-10

Few shot classification of auroral vortices with space physics parameters

- 1.
  School of Electronic Engineering, Xidian University, Xi'an 710071, China
- 2.
  MNR Key Laboratory for Polar Science, Polar Research Institute of China, Shanghai 200136, China
- 3.
  Ocean College, Zhejiang University, Zhoushan Zhejiang 316021, China
- 4.
  Institute of Applied Meteorology, Beijing 100029, China

More Information

Corresponding author: HU ZeJun, E-mail: huzejun@pric.org.cn

Received Date: 17 August 2022

Revised Date: 11 April 2023

Available Online: 10 November 2023

摘要

摘要:
复杂的空间物理过程导致极光出现各类复杂的形态.极光形态的自动分类有助于分析极光发生机制以及了解空间物理过程.当前的极光分类主要采用传统的机器学习或深度学习的方法, 需要大数据作为支撑.但在实际发生的极光事件中, 一些事件相比其他极光事件出现频率小(如极光涡旋), 这样的事件是小样本事件.本文提出一种基于注意力机制, 结合空间物理参数信息的小样本极光事件分类方法.从北极黄河站2003—2017年全天空极光数据集中, 构建了一个有85个极光涡旋序列的实验数据集.利用该方法, 对极光涡旋这一类特殊小样本极光事件进行分类研究.研究结果显示, 加入空间物理参数有助于极光涡旋的分类, 分类准确率从56.37%提升到66.25%, 同时也表明空间环境参数对极光涡旋的产生有着明显的调制作用.
- 极光涡旋 /
- 空间参数 /
- 深度学习 /
- 小样本学习 /
- 多模态信息
Abstract:
The complex morphological characteristics of auroras are closely related with the complex processes of space physics. The automatic classification of auroral morphology helps us to analyze the mechanism of auroral formation and understand the processes of space physics. Currently, traditional machine learning or deep learning methods are mainly adopted in the auroral classification, which need to be supported by large data. However, some specific aurora events appear less frequently, such as aurora vortices, and can be considered as small sample events compared with other aurora events. In this paper, it presents a few shot learning methods for aurora event classification based on attention mechanism and space physics parameter information, and a dataset with 85 auroral vortices events selected from 2003—2017 all-sky aurora dataset of the Chinese Yellow River Station in Arctic is constructed. The results show that space physics parameters is helpful for auroral vortices classification, the accuracy increases from 56.37% to 66.25%. In addition, the space environment parameters have obvious modulation effect on the occurrence of auroral vortices.
- Auroral vortices /
- Space parameter /
- Deep learning /
- Few shot learning /
- Multimode information

HTML全文

图 1

不同类型的涡旋结构示例

Figure 1.

Different auroral vortex structures

下载: 全尺寸图片幻灯片

图 2

涡旋事件时间分布图

Figure 2.

Temporal distribution of auroral vortices

下载: 全尺寸图片幻灯片

图 3

算法流程图

Figure 3.

The flowchart of our proposed algorithm structure

下载: 全尺寸图片幻灯片

图 4

注意力模块结构(其中m，h，w分别代表特征通道数、高度和宽度)

Figure 4.

Attention module structure

下载: 全尺寸图片幻灯片

图 5

多模态信息融合的流程

Figure 5.

The flowchart of multimodal information fusion

下载: 全尺寸图片幻灯片

图 6

添加不同时间段物理参数实验结果

Figure 6.

Experimental results with different physics parameters in different time periods

下载: 全尺寸图片幻灯片

图 7

添加不同时刻物理参数实验结果

Figure 7.

Experimental results with different physics parameters in different times

下载: 全尺寸图片幻灯片

表 1

不同小样本模型实验结果

Table 1.

Experimental results of different few shot models

方法	网络结构	空间参数	准确率
DN4	Conv1-2-3-4	√	42.5%
DN4	Conv1-2-3-4	×	40%
Ours	ResNet-12	√	66.25%
	ResNet-12	×	56.37%
	ResNet-34	×	52.5%%
	ResNet-50	×	44.75%
	ResNet-101	×	37%

下载: 导出CSV

表 2

最高准确率下的混淆矩阵

Table 2.

Confusion matrix with the highest accuracy

真实类别	预测类别
真实类别	A	B	C	D
A	44	18	8	10
B	11	69	0	0
C	25	0	39	16
D	8	3	9	60
注: A, B, C, D分别代表大扭曲结构的极光弧，大扭曲结构的极光射线带，极光弧涡旋，射线簇涡旋，每一行为真实类别，每一列为预测类别.

下载: 导出CSV

表 3

不同空间参数组合对分类准确率影响的实验结果

Table 3.

Experimental results with different parameter combinations

No.	输入空间参数						准确率
No.	行星际磁场			太阳风参数		亚暴指数	准确率
1	B_x	B_y	B_z	V_p	N_p	AE	66.25%
2	B_x						62.19%
3		B_y					62.19%
4			B_z				63.75%
5				V_p			62.19%
6					N_p		58.13%
7						AE	59.69%
8	B_x	B_y					58.13%
8	B_x		B_z				58.13%
10	B_x			V_p			60.94%
11	B_x				N_p		61.25%
12	B_x					AE	59.06%
13		B_y	B_z				59.69%
14		B_y		V_p			60.62%
15		B_y			N_p		57.81%
16		B_y				AE	58.75%
17			B_z	V_p			61.25%
18			B_z		N_p		58.44%
19			B_z			AE	58.13%
20				V_p	N_p		57.81%
21				V_p		AE	61.56%
22					N_p	AE	58.44%
23	B_x	B_y	B_z				63.12%
24	B_x	B_y		V_p			59.06%
25	B_x	B_y			N_p		60.31%
26	B_x	B_y				AE	60.31%
27	B_x		B_z	V_p			56.88%
28	B_x		B_z		N_p		62.5%
29	B_x		B_z			AE	60.94%
30	B_x			V_p	N_p		58.44%
31	B_x			V_p		AE	58.44%
32	B_x				N_p	AE	58.13%
33		B_y	B_z	V_p			59.06%
34		B_y	B_z		N_p		57.81%
35		B_y	B_z			AE	58.13%
36		B_y		V_p	N_p		59.06%
37		B_y		V_p		AE	59.06%
38		B_y			N_p	AE	56.56%
39			B_z	V_p	N_p		55.62%
40			B_z	V_p		AE	57.19%
41			B_z		N_p	AE	56.56%
42				V_p	N_p	AE	59.06%
43	B_x	B_y	B_z	V_p			61.56%
44	B_x	B_y	B_z		N_p		56.26%
45	B_x	B_y	B_z			AE	61.88%
46	B_x	B_y		V_p	N_p		58.44%
47	B_x	B_y		V_p		AE	61.88%
48	B_x	B_y			N_p	AE	58.44%
49	B_x		B_z	V_p	N_p		61.25%
50	B_x		B_z	V_p		AE	57.5%
51	B_x		B_z		N_p	AE	59.69%
52	B_x			V_p	N_p	AE	61.88%
53		B_y	B_z	V_p	N_p		57.5%
54		B_y	B_z	V_p		AE	61.56%
55		B_y	B_z		N_p	AE	56.56%
56		B_y		V_p	N_p	AE	57.5%
57			B_z	V_p	N_p	AE	60.31%
58	B_x	B_y	B_z	V_p	N_p		62.19%
59	B_x	B_y	B_z	V_p		AE	62.81%
60	B_x	B_y	B_z		N_p	AE	59.06%
61	B_x	B_y		V_p	N_p	AE	57.81%
62	B_x		B_z	V_p	N_p	AE	60.94%
63		B_y	B_z	V_p	N_p	AE	60.31%

下载: 导出CSV

表 4

添加Dst指数的实验结果

Table 4.

Experimental results with Dst parameter and other parameters

No.	输入空间参数							准确率
No.	行星际磁场			太阳风参数		亚暴指数	地磁指数	准确率
1	B_x	B_y	B_z	V_p	N_p	AE	Dst	56.56%
2	B_x	B_y	B_z				Dst	62.19%
3				V_p			Dst	58.44%
4					N_p		Dst	59.38%
5						AE	Dst	60.62%
6	B_x	B_y	B_z	V_p			Dst	61.25%
7	B_x	B_y	B_z		N_p		Dst	58.44%
8	B_x	B_y	B_z			AE	Dst	58.44%
9				V_p	N_p		Dst	56.25%
10				V_p		AE	Dst	60.31%
11					N_p	AE	Dst	60.31%
12	B_x	B_y	B_z	V_p	N_p		Dst	61.56%
13	B_x	B_y	B_z	V_p		AE	Dst	57.19%
14	B_x	B_y	B_z		N_p	AE	Dst	60.31%
15				V_p	N_p	AE	Dst	60.31%

下载: 导出CSV

参考文献(63)

	Andrychowicz M, Denil M, Colmenarejo S G, et al. 2016. Learning to learn by gradient descent by gradient descent. //Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona, Spain: Curran Associates Inc., 3988-3996.
	Balan N, Zhang Q H, Shiokawa K, et al. 2019. IpsDst of Dst storms applied to ionosphere-thermosphere storms and low-latitude aurora. Journal of Geophysical Research: Space Physics, 124(11): 9552-9565, doi: 10.1029/2019ja027080.
	Bertinetto L, Henriques J F, Valmadre J, et al. 2016. Learning feed-forward one-shot learners. //Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona, Spain: Curran Associates Inc., 523-531.
	Cai Q, Pan Y W, Yao T, et al. 2018. Memory matching networks for one-shot image recognition. //Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 4080-4088.
	Chen X C, Han D S, Lorentzen D A, et al. 2017. Dynamic properties of throat aurora revealed by simultaneous ground and satellite observations. Journal of Geophysical Research: Space Physics, 122(3): 3469-3486, doi: 10.1002/2016JA023033.
	Clausen L B N, Nickisch H. 2018. Automatic classification of auroral images from the Oslo Auroral THEMIS (OATH) data set using machine learning. Journal of Geophysical Research: Space Physics, 123(7): 5640-5647, doi: 10.1029/2018JA025274.
	Davis T N, Hallinan T J. 1976. Auroral spirals, 1. Observations. Journal of Geophysical Research, 81(22): 3953-3958, doi: 10.1029/JA081i022p03953.
	Fasel G J. 1995. Dayside poleward moving auroral forms: A statistical study. Journal of Geophysical Research: Space Physics, 100(A7): 11891-11905, doi: 10.1029/95JA00854.
	Hallinan T J. 1976. Auroral spirals, 2. Theory. Journal of Geophysical Research, 81(22): 3959-3965, doi: 10.1029/JA081i022p03959.
	Han B, Zhao X J, Tao D C, et al. 2014. Dayside aurora classification via BIFs-based sparse representation using manifold learning. International Journal of Computer Mathematics, 91(11): 2415-2426, doi: 10.1080/00207160.2013.831084.
	Han B, Gao X B, Liu H, et al. 2015. Auroral oval boundary modeling based on deep learning method. //5th International Conference on Intelligent Science and Big Data Engineering. Suzhou, China: Springer, 96-106, doi: 10.1007/978-3-319-23862-3_10.
	Han B, Song Y T, Gao X B, et al. 2016. Dynamic aurora sequence recognition using Volume Local Directional Pattern with local and global features. Neurocomputing, 184: 168-175, doi: 10.1016/j.neucom.2015.07.126.
	Han B, Chu F Y, Gao X B, et al. 2017a. A multi-size kernels CNN with eye movement guided task-specific initialization for aurora image classification. //2nd CCF Chinese Conference on Computer Vision. Tianjin, China: Springer, 533-544, doi: 10.1007/978-981-10-7302-1_44.
	Han B, Jia Z H, Gao X B. 2017b. Improved PCANet for aurora images classification. Journal of Xidian University (in Chinese), 44(1): 83-88.
	Han B, Lian H, Hu Z J, et al. 2019a. Modeling of ultraviolet auroral oval intensity based on Generative Adversarial Networks. //AGU Fall Meeting Abstracts. AGU.
	Han B, Lian H F, Hu Z J. 2019b. Modeling of ultraviolet auroral oval boundaries based on neural network technology. Scientia Sinica Technologica (in Chinese), 49: 531-542, doi: 10.1360/N092018-00227.
	Han D S, Chen X C, Liu J J, et al. 2015. An extensive survey of dayside diffuse aurora based on optical observations at Yellow River Station. Journal of Geophysical Research: Space Physics, 120(9): 7447-7465, doi: 10.1002/2015JA021699.
	Han D S, Nishimura Y, Lyons L R, et al. 2016. Throat aurora: Theionospheric signature of magnetosheath particles penetrating into the magnetosphere. Geophysical Research Letters, 43(5): 1819-1827, doi: 10.1002/2016GL068181.
	Han D S, Li J X, Nishimura Y, et al. 2017. Coordinated observations of two types of diffuse auroras near magnetic local noon by Magnetospheric Multiscale mission and ground all-sky camera. Geophysical Research Letters, 44(16): 8130-8139, doi: 10.1002/2017GL074447.
	Han D S, Xiong Y T, Shi R, et al. 2023. A unified model of cusp spot, High Latitude Dayside aurora (HiLDA)/(Space Hurricane), and 15MLT-PCA. Earth and Planetary Physics, 7(4): 513-519, doi: 10.26464/epp2023046.
	Han Y Y, Han B, Hu Z J, et al. 2020. Prediction and variation of the auroral oval boundary based on a deep learning model and space physical parameters. Nonlinear Processes in Geophysics, 27(1): 11-22, doi: 10.5194/npg-27-11-2020.
	Han Y Y, Han B, Gao X B. 2023. Auroral substorm event recognition method combining eye movement information and sequence fingerprint. Journal of Frontiers of Computer Science and Technology (in Chinese), 17(1): 187-197, doi: 10.3778/j.issn.1673-9418.2105044.
	He K M, Zhang X Y, Ren S Q, et al. 2016. Deep residual learning for image recognition. //Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 770-778.
	Hiraki Y. 2016. Ionospheric current system accompanied by auroral vortex streets. arXiv reprint arXiv: 1603.09075.
	Hu Z J, Yang H G, Huang D S, et al. 2009. Synoptic distribution of dayside aurora: Multiple-wavelength all-sky observation at Yellow River Station in Ny-Ålesund, Svalbard. Journal of Atmospheric and Solar-Terrestrial Physics, 71(8-9): 794-804, doi: 10.1016/j.jastp.2009.02.010.
	Hu Z J, Yang H G, Liang J, et al. 2010. The 4-emission-core structure of dayside aurora oval observed by all-sky imager at 557.7 nm in Ny-Ålesund, Svalbard. Journal of Atmospheric and Solar-Terrestrial Physics, 72(7-8): 638-642, doi: 10.1016/j.jastp.2010.03.005.
	Hu Z J, Yang H G, Han D S, et al. 2012. Daysideauroral emissions controlled by IMF: A survey for dayside auroral excitation at 557.7 and 630.0 nm in Ny-Ålesund, Svalbard. Journal of Geophysical Research: Space Physics, 117(A2): A02201, doi: 10.1029/2011JA017188.
	Hu Z J, Yang H G, Hu H Q, et al. 2013. The hemispheric conjugate observation of postnoon "bright spots" /auroral spirals. Journal of Geophysical Research: Space Physics, 118(4): 1428-1434, doi: 10.1002/jgra.50243.
	Hu Z J, Yang Q J, Liang J M, et al. 2017a. Variation and modeling of ultraviolet auroral oval boundaries associated with interplanetary and geomagnetic parameters. Space Weather, 15(4): 606-622, doi: 10.1002/2016SW001530.
	Hu Z J, Yang H G, Ebihara Y, et al. 2017b. Surveys of 557.7/630.0 nm dayside auroral emissions in Ny-Ålesund, Svalbard, and South Pole Station. //Haaland S, Runov A, Forsyth C eds. Dawn-Dusk Asymmetries in Planetary Plasma Environments. American Geophysical Union, 143-154, doi: 10.1002/9781119216346.ch11.
	Hu Z J, Han B, Lian H F. 2020. Modeling of ultraviolet auroral intensity based on Generalized Regression Neural Network associated with IMF/solar wind and geomagnetic parameters. Chinese Journal of Geophysics (in Chinese), 63(5): 1738-1750, doi: 10.6038/cjg2020N0151.
	Hu Z J, Han B, Zhang Y S, et al. 2021. Modeling of ultraviolet aurora intensity associated with interplanetary and geomagnetic parameters based on neural networks. Space Weather, 19(11): e2021SW002751, doi: 10.1029/2021SW002751.
	Jia Z H, Han B, Gao X B. 2015. 2DPCANet: dayside aurora classification based on deep learning. //CCF Chinese Conference on Computer Vision. Xi′an, China: Springer, 323-334, doi: 10.1007/978-3-662-48570-5_32.
	Keiling A, Angelopoulos V, Runov A, et al. 2009. Substorm current wedge driven by plasma flow vortices: THEMIS observations. Journal of Geophysical Research: Space Physics, 114(A1): A00C22, doi: 10.1029/2009JA014114.
	Kvammen A, Wickstrøm K, McKay D, et al. 2020. Auroral image classification with deep neural networks. Journal of Geophysical Research: Space Physics, 125(10): e2020JA027808, doi: 10.1029/2020JA027808.
	Li W B, Wang L, Xu J L, et al. 2019. Revisiting local descriptor based image-to-class measure for few-shot learning. //Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: IEEE, 7253-7260.
	Liu Y Y, Su Y T, Liu A A, et al. 2020. LCC: learning to customize and combine neural networks for few-shot learning. arXiv preprint arXiv: 1904.08479.
	Meinel A B. 1954. The morphology of the aurora. Proceedings of the National Academy of Sciences of the United States of America, 40(10): 943-950, doi: 10.1073/pnas.40.10.943.
	Munkhdalai T, Yu H. 2017. Meta networks. //Proceedings of the 34th International Conference on Machine Learning. Sydney, NSW, Australia: JMLR. org, 2554-2563.
	Munkhdalai T, Yuan X D, Mehri S, et al. 2018. Rapid adaptation with conditionally shifted neurons. //Proceedings of the 35th International Conference on Machine Learning. Stockholm, Sweden: PMLR, 3664-3673.
	Naik D K, Mammone R J. 1992. Meta-neural networks that learn by learning. //IJCNN International Joint Conference on Neural Networks. Baltimore, MD, USA: IEEE, 437-442.
	Ohma A, Madelaire M, Laundal K M, et al. 2023. Background removal from global auroral images: Data driven dayglow modelling. Earth and Planetary Physics, 7(6): 1-11, doi: 10.26464/epp2023051.
	Oreshkin B N, Rodríguez P, Lacoste A. 2018. TADAM: Task dependent adaptive metric for improved few-shot learning. //Proceedings of the 32nd International Conference on Neural Information Processing Systems. Montréal, Canada: Curran Associates Inc., 719-729.
	Ravi S, Larochelle H. 2017. Optimization as a model for few-shot learning. //5th International Conference on Learning Representations. Toulon, France.
	Shi R, Han D, Ni B, et al. 2012. Intensification of dayside diffuseauroral precipitation: contribution of dayside Whistler-mode chorus waves in realistic magnetic fields. Annales Geophysicae, 30(9): 1297-1307, doi: 10.5194/angeo-30-1297-2012.
	Shi R, Hu Z J, Ni B B, et al. 2014. Modulation of the dayside diffuse auroral intensity by the solar wind dynamic pressure. Journal of Geophysical Research: Space Physics, 119(12): 10092-10099, doi: 10.1002/2014JA020180.
	Snell J, Swersky K, Zemel R. 2017. Prototypical networks for few-shot learning. //Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California, USA: Curran Associates Inc., 4080-4090.
	Song Y T, Han B, Gao X B. 2016. Tensor based dynamic textures model for aurora sequences classification. Journal of Nanjing University (Natural Sciences) (in Chinese), 52(1): 184-193, doi: 10.13232/j.cnki.jnju.2016.01.021.
	Sung F, Yang Y X, Zhang L, et al. 2018. Learning to compare: Relation network for few-shot learning. //Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 1199-1208.
	Thrun S, Pratt L. 1998. Learning to learn: Introduction and overview. //Learning to Learn. New York, NY: Springer, 3-17.
	Vinyals O, Blundell C, Lillicrap T, et al. 2016. Matching networks for one shot learning. //Proceedings of the 30th International Conference on Advances in Neural Information Processing Systems. Barcelona, Spain: Curran Associates Inc., 3637-3645.
	Wang Q, Liang J M, Hu Z J, et al. 2010. Spatial texture based automatic classification of dayside aurora in all-sky images. Journal of Atmospheric and Solar-Terrestrial Physics, 72(5-6): 498-508, doi: 10.1016/j.jastp.2010.01.011.
	Yan Y. 2018. Study on theclassification and physical characteristic of day-side aurora [Master′s thesis] (in Chinese). Xi′an: Xidian University.
	Yang Q J, Hu Z J. 2018. A comparative study of auroral morphology distribution between the Northern and Southern Hemisphere based on automatic classification. Geoscientific Instrumentation, Methods and Data Systems, 7(1): 113-122, doi: 10.5194/gi-7-113-2018.
	Yang Q J, Zhou P H. 2020. Representation and classification of auroral images based on convolutional neural networks. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 13: 523-534. doi: 10.1109/JSTARS.2020.2969245
	Yang X, Gao X B, Song B, et al. 2018. Aurora image search with contextual CNN feature. Neurocomputing, 281: 67-77, doi: 10.1016/j.neucom.2017.11.059.
	Zhou D Y, Bousquet O, Lal T N, et al. 2003. Learning with local and global consistency. //Proceedings of the 16th International Conference on Neural Information Processing Systems. Whistler, British Columbia, Canada: MIT Press, 321-328.
	韩冰, 贾中华, 高新波. 2017. 改进的主成分分析网络极光图像分类方法. 西安电子科技大学学报(自然科学版), 44(1): 83-88. https://www.cnki.com.cn/Article/CJFDTOTAL-XDKD201701015.htm
	韩冰, 连慧芳, 胡泽骏. 2019. 基于神经网络模型的紫外极光卵边界建模. 中国科学: 技术科学, 49(5): 531-542, doi: 10.1360/N092018-00227.
	韩怡园, 韩冰, 高新波. 2023. 结合眼动信息和序列指纹的亚暴事件识别方法. 计算机科学与探索, 17(1): 187-197, doi: 10.3778/j.issn.1673-9418.2105044.
	胡泽骏, 韩冰, 连慧芳. 2020. 基于广义回归神经网络的行星际/太阳风参数和地磁指数的紫外极光强度建模. 地球物理学报, 63(5): 1738-1750, doi: 10.6038/cjg2020N0151.
	宋亚婷, 韩冰, 高新波. 2016. 基于张量动态纹理模型的极光视频分类. 南京大学学报(自然科学), 52(1): 184-193, doi: 10.13232/j.cnki.jnju.2016.01.021.
	严月. 2018. 日侧极光分类方法及物理特征研究[硕士论文]. 西安: 西安电子科技大学.

施引文献

资源附件(0)

访问统计

图(7)

表(4)

计量

文章访问数:
PDF下载数:
施引文献: 0

融合空间物理参数的小样本极光涡旋自动分类研究

作者简介: 韩冰, 女, 1978年生, 教授, 主要从事信号处理、视觉感知、计算机视觉、模式识别等方向研究.E-mail: bhan@xidian.edu.cn

通讯作者: 胡泽骏, 男, 1978年生, 研究员, 博士生导师, 研究方向为极光物理, 极区太阳风-磁层-电离层耦合, 极区空间环境/空间天气方面的研究.E-mail: huzejun@pric.org.cn

Few shot classification of auroral vortices with space physics parameters

Corresponding author: HU ZeJun, E-mail: huzejun@pric.org.cn

计量

出版历程

目录

作者简介:
韩冰, 女, 1978年生, 教授, 主要从事信号处理、视觉感知、计算机视觉、模式识别等方向研究.E-mail: bhan@xidian.edu.cn