基于深度学习的云参量反演方法研究

摘要/Abstract

摘要： 有效的云检测与云相态判识对于农业、气候及人类生活具有重要意义, 而这些数据的获取离不开卫星遥感。卫星遥感数据在当今社会的生产和生活中都扮演着至关重要的角色, 众多领域的发展都离不开卫星遥感数据的支持。随着高精度传感器的发展, 传统研究方法无法满足大规模、高维度数据的高效挖掘与处理, 因此深度学习技术在遥感领域得到了快速的发展。基于深度学习技术提出了一种结合多波段遥感影像的云检测及云相态判识的方法。该方法采用 MODIS 云产品影像作为样本, 将不同波段信息作为特征值, 分别建立针对云检测与云相态判识研究任务的多个数据库, 并采用 DeepLab V3+ 模型进行训练并预测, 从而完成高精度的云检测及云相态判识任务。与传统方法相比, 该方法高效便捷、特征提取能力较强, 将多波段作为特征值输入模型进行预测时, 该方法展现了良好的结果。

关键词: 中分辨率成像光谱仪, Deeplab V3+, 云检测, 云相态

Abstract: Effective cloud detection and cloud phase identification are of great significance to agriculture, climate and human life, and the acquisition of these data is mainly from satellite remote sensing. Satellite remote sensing data plays a vital role in the production and life of current society, and the development of many fields is inseparable from the support of satellite remote sensing data. With the development of high-precision sensors, traditional research methods cannot meet the requirements of large-scale and high-dimensional data mining and processing, so deep learning technology has been rapidly developed in the field of remote sensing. Based on deep learning technology, a method of cloud detection and cloud phase recognition combined with multi-band remote sensing images is proposed in this work. MODIS cloud product images are used as samples, the different waveband information is used as eigenvalue to establish multiple databases for cloud detection and cloud phase state recognition research tasks, and then Deeplab V3+ model is used for training and prediction, so as to complete the high-precision cloud detection and cloud phase state recognition. Compared with the traditional methods, the proposed method is more efficient and convenient, and has stronger feature extraction ability. When multi-band is used as the eigenvalue input model for prediction, the method shows good performance.

Key words: moderate-resolution imaging spectroradiometer, Deeplab V3+, cloud detection, cloud phase identification

中图分类号:

X831

吴文涵, 麻金继∗, 孙二昌, 郭金雨, 杨光, 王宇瑶. 基于深度学习的云参量反演方法研究[J]. 大气与环境光学学报, 2022, 17(4): 453-464.

WU Wenhan, MA Jinji∗, SUN Erchang, GUO Jinyu, YANG Guang, WANG Yuyao. Research on cloud parameter inversion method based on deep learning[J]. Journal of Atmospheric and Environmental Optics, 2022, 17(4): 453-464.

参考文献 29

[1]	Rossow W B, Mosher F, Kinsella E, et al. ISCCP cloud algorithm intercomparison [J]. Journal of Climate and Applied
	Meteorology, 1985, 24(9): 877-903.
[2]	Yang J, Chen B J, Yin Y. Physics of Clouds and Precipitation [M]. Beijing: Meteorological Press, 2011.
	杨军, 陈宝君, 银燕. 云降水物理学 [M]. 北京: 气象出版社, 2011.
[3]	Li W B. Atmospheric Remote Sensing [M]. Beijing: Peking University Press, 2014.
	李万彪. 大气遥感 [M]. 北京: 北京大学出版社, 2014.
[4]	Strabala K I, Ackerman S A, Menzel W P. Cloud properties inferred from 8-12-µm data [J]. Journal of Applied Meteorology,
19	94, 33(2): 212-229.
[5]	Baum B A, Soulen P F, Strabala K I, et al. Remote sensing of cloud properties using MODIS airborne simulator imagery during
	SUCCESS: 2. Cloud thermodynamic phase [J]. Journal of Geophysical Research: Atmospheres, 2000, 105(D9): 11781-11792.
[6]	Segal-Rozenhaimer M, Li A L, Das K, et al. Cloud detection algorithm for multi-modal satellite imagery using convolutional
	neural-networks (CNN) [J]. Remote Sensing of Environment, 2020, 237: 111446.
[7]	He Y M, Wang H J, Guan H. Application of support vector machines in cloud detection using EOS/MODIS [C]. International
	Symposium on Photoelectronic Detection and Imaging 2009: Advances in Infrared Imaging and Applications, 2009, 7383:
20	9-216.
[8]	Yang C P, Guo J. Improved cloud phase retrieval approaches for China′s FY-3A/VIRR multi-channel data using artificial neural
	networks [J]. Optik, 2016, 127(4): 1797-1803.
[9]	Hinton G E, Salakhutdinov R R. Reducing the dimensionality of data with neural networks [J]. Science, 2006, 313(5786):
50	4-507.
[10]	Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE,
19	98, 86(11): 2278-2324.
[11]	Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks [J]. Communications
	of the ACM, 2017, 60(6): 84-90.
[12]	He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition [C]. 2016 IEEE Conference on Computer
	Vision and Pattern Recognition. June 27-30, 2016, Las Vegas, NV, USA. IEEE, 2016: 770-778.
[13]	Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation [C]. Medical Image
	Computing and Computer-Assisted Intervention-MICCAI 2015, 2015: 234-241.
[14]	Altini N, Cascarano G D, Brunetti A, et al. Semantic segmentation framework for glomeruli detection and classification in
	kidney histological sections [J]. Electronics, 2020, 9(3): 503.
[15]	Hu X L, Liu Z Y, Zhou H Y, et al. Deep HT: A deep neural network for diagnose on MR images of tumors of the hand [J].
	PLoS One, 2020, 15(8): e0237606.
[16]	Huang X D, Zhang H, Zhuo L, et al. TISNet-enhanced fully convolutional network with encoder-decoder structure for tongue
	image segmentation in traditional Chinese medicine [J]. Computational and Mathematical Methods in Medicine, 2020, 2020:
60	29258.
[17]	Guo H X, He G J, Jiang W, et al. A multi-scale water extraction convolutional neural network (MWEN) method for GaoFen-1
	remote sensing images [J]. ISPRS International Journal of Geo-Information, 2020, 9(4): 189.
[18]	Wang G J, Wu M J, Wei X K, et al. Water identification from high-resolution remote sensing images based on multidimensional
	densely connected convolutional neural networks [J]. Remote Sensing, 2020, 12(5): 795.
[19]	Liu P H, Liu X P, Liu M X, et al. Building footprint extraction from high-resolution images via spatial residual inception
	convolutional neural network [J]. Remote Sensing, 2019, 11(7): 830.
[20]	Shi Q, Liu M X, Liu X P, et al. Domain adaption for fine-grained urban village extraction from satellite images [J]. IEEE
	Geoscience and Remote Sensing Letters, 2020, 17(8): 1430-1434.
[21]	Wang Y H, Liang B X, Ding M, et al. Dense semantic labeling with atrous spatial pyramid pooling and decoder for highresolution remote sensing imagery [J]. Remote Sensing, 2018, 11(1): 20.
[22]	Yao X D, Yang H, Wu Y L, et al. Land use classification of the deep convolutional neural network method reducing the loss of
	spatial features [J]. Sensors (Basel), 2019, 19(12): 2792.
[23]	Jiang N, Li J Y. An improved semantic segmentation method for remote sensing images based on neural network [J]. Traitement
	Du Signal, 2020, 37(2): 271-278.
[24]	Zhan Z Q, Zhang X M, Liu Y, et al. Vegetation land use/land cover extraction from high-resolution satellite images based on
	adaptive context inference [J]. IEEE Access, 2020, 8: 21036-21051.
[25]	Venugopal N. Automatic semantic segmentation with DeepLab dilated learning network for change detection in remote sensing
	images [J]. Neural Processing Letters, 2020, 51(3): 2355-2377.
[26]	Chen L C, Zhu Y K, Papandreou G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation
[C]	Computer Vision-ECCV 2018, 2018, 833-851.
[27]	He K M, Zhang X Y, Ren S Q, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition [J]. IEEE
	Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916.
[28]	Ben-Dor E. A precaution regarding cirrus cloud detection from airborne imaging spectrometer data using the 1.38 µm water
	vapor band [J]. Remote Sensing of Environment, 1994, 50(3): 346-350.
[29]	Wang L Y. Research on Semantic Segmentation of Remote Sensing Images of the Ground Objects Based on Deeplab V3+
	Network [D]. Harbin: Harbin Institute of Technology, 2020.
	王蓝玉. 基于 Deeplab V3+ 网络的遥感地物图像语义分割研究 [D]. 哈尔滨: 哈尔滨工业大学, 2020。