注意力机制下多尺度特征融合生成对抗网络的白天海雾监测方法
Daytime sea fog detection based on multi-scale feature fusion of generate adversarial network under attention mechanism
- 2022年 页码：1-12
DOI： 10.11834/jrs.20221621
引用
扫 描 看 全 文
摘要
海雾是海上一种常见的天气现象，它使能见度降低，给海上交通和作业带来极大威胁。传统的卫星遥感海雾监测算法在准确率、可移植性及自动化程度等方面都有待改善。本文在注意力机制下，利用卫星遥感云图，提出一种多尺度特征融合生成对抗网络的白天海雾监测方法。该方法引入通道注意力机制，通过学习不同输入通道的权重，提升了网络对于重要通道云图的关注度；在此基础上，采用多尺度特征融合以获取海雾的多尺度信息，使提取的特征能兼顾海雾的整体及细节特性；为了进一步提高算法对于对海雾边缘的界定能力，本文引入对抗网络对海雾监测的生成网络进行监督，从而得到更精细的海雾区域。在测试云图的海雾监测实验中，命中率（POD）、临界成功指数（CSI）及误报率（FAR）分别为90.5%、81.28%和10.86%，均优于传统海雾监测方法以及其它基于深度学习的方法，这表明本文方法可以有效提升海雾监测的精度，研究成果对于海上船只航行、渔业生产、国防军事等具有重要意义。
Abstract
Sea fog is a common weather phenomenon at sea. It will reduce visibility at sea and pose a significant threat to maritime traffic and other operations. Traditional sea fog detection algorithms using satellite remote sensing have low accuracy, poor portability, and low automation. Although some existing deep learning-based sea fog monitoring algorithms have been improved, they do not consider the spectral characteristics of sea fog in different channels, and the accuracy of sea fog monitoring is not high, especially in edge recognition.In order to improve the accuracy of sea fog detection, a daytime sea fog detection method was proposed, which based on multi-scale feature fusion of generate adversarial network (GAN) under attention mechanism. Firstly, according to the spectral response of sea fog in different imaging channels of meteorological satellite, the satellite cloud images of different imaging channels that can reflect the characteristics of sea fog are selected as the input of the network. Meanwhile, in order to make the network better focus on the significant imaging channels under multi-channel input, the channel attention mechanism is introduced to measure the weights of different input channels. Then, in order to solve the problem of the loss detail features of the cloud image caused by the pooling operation of the traditional deep network, a multi-scale feature fusion mechanism is adopted to fuse the feature maps of different levels of the network to obtain the multi-scale features of the sea fog. Finally, in view of the difficulty of traditional methods to accurately describe the edge of sea fog, the generation network for sea fog detection is supervised by adversarial network, so as to accurately define the edge of sea fog and reduce the false alarm rate.This article takes the Yellow Sea and the Bohai Sea (,,http://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=30727721&type=,http://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=30727720&type=,22.60600090,2.45533323,，,,http://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=30727799&type=,http://html.publish.founderss.cn/rc-pub/api/common/picture?pictureId=30727722&type=,16.08666801,2.53999996,) as the research area. Since March to June each year is the period of high incidence of sea fog in the Yellow Sea and the Bohai Sea, we produce a data set based on the weather satellite monitoring report of the National Meteorological Center from March to June 2017-2020. After the model training in this article, in terms of quantitative indicators of sea fog detection, probability of detection (POD), critical success index (CSI) and false positive rate (FAR) of our method are 90.5%, 81.28% and 10.86% respectively, which are better than other methods.The experimental results show that the method in this paper can effectively improve the accuracy of sea fog identification, which is of great significance for marine vessel navigation, fishery production, national defense and military affairs.
关键词
海雾监测卫星遥感注意力机制生成对抗网络多尺度特征融合
Keywords
Sea fog monitoringRemote sensingAttention mechanismGenerate adversarial networkMulti-scale feature fusion
references
Bendix, J，Thies, Boris，Cermak, Jan， Nau, Thomas. 2005. Ground Fog Detection from Space Based on MODIS Daytime Data—A Feasibility Study. Weather and Forecasting, 20(6):989-1005. [DOI 10.1175/WAF886.1http://dx.doi.org/10.1175/WAF886.1]
Chen L. 2006. The data processing method of EOS/MODIS data and its application in daytime low cloud and heavy fog monitoring. Nanjing University of Information Science and Technology.
陈林. 2006. EOS/MODIS资料的数据处理方法及其在白天低云大雾监测中的应用. 南京信息工程大学.[DOI 10.7666/d.y863542http://dx.doi.org/10.7666/d.y863542]
Cermak J , Bendix J . 2008. A novel approach to fog/low stratus detection using Meteosat 8 data. Atmospheric Research, 87(3-4):279-292. [DOI 10.1016/j.atmosres.2007.11.009http://dx.doi.org/10.1016/j.atmosres.2007.11.009]
Deng Y J, Wang J C, Cao J. Monitoring daytime fog in the South China Sea based on MODIS remote sensing data. Journal of Tropical Meteorology, 2013(6).
邓玉娇, 王捷纯, 曹静. 基于MODIS遥感资料监测南海白天雾. 热带气象学报, 2013(6). [DOI CNKI:SUN:RDQX.0.2013-06-019]
Eyre J R , Brownscombe J L , Allan R J . Detection of fog at night using Advanced Very High Resolution Radiometer (AVHRR)imagery. meteorological magazine. [DOI 10.1007/ s00253-006-0446-y]
Goodfellow I J, Pouget-Abadie J, Mirza M, et al. 2014. Generative Adversarial Networks[J]. Advances in Neural Information Processing Systems, 3:2672-2680. [DOI 10.1145/3422622http://dx.doi.org/10.1145/3422622]
Hunt G E . 1973. Radiative properties of terrestial clouds at visible and infra-red thermal window wavelengths. Quarterly Journal of the Royal Meteorological Society, 99(420):346-369. [DOI 10.1002/qj.49709942013http://dx.doi.org/10.1002/qj.49709942013]
Huang H, Lin L, Tong R, et al. 2020. UNet 3+: A Full-Scale Connected UNet for Medical Image Segmentation// ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE. [DOI 10.1109/ICASSP40776.2020.9053405http://dx.doi.org/10.1109/ICASSP40776.2020.9053405]
Ishida H , Miura K , Matsuda T , et al. , 2014. Scheme for detection of low clouds from geostationary weather satellite imagery. Atmospheric Research, 143(jun.):250-264. [DOI 10.1016/j.atmosres.2014.02.015http://dx.doi.org/10.1016/j.atmosres.2014.02.015]
Li Y C, Sun H, Li X G, Tang Y, Xu M. 2001. Research on monitoring daytime fog with GMS-5 meteorological satellite data remote sensing. Chinese Journal of Atmospheric Sciences, 24(3):343-349.
李亚春，孙涵，李湘阁，唐勇，徐萌. 2001. 用GMS-5气象卫星资料遥感监测白天雾的研究. 大气科学学报,24(3):343-349. DOI 10.3969/j.issn.1674-7097.2001.03.007http://dx.doi.org/10.3969/j.issn.1674-7097.2001.03.007
Liu S X, Yi L, Zhang S P, Shi X Z, Xue Y C. Research on the Daytime Yellow Sea Sea Fog Satellite Inversion Based on the Fully Convolutional Neural Network Method. Ocean Limnology Bulletin, 2019(6): 13-22.
刘树霄，衣立，张苏平，时晓曚，薛允传. 基于全卷积神经网络方法的日间黄海海雾卫星反演研究. 海洋湖沼通报, 2019(6):13-22.
Liu Y J. 2019. Research on Image Semantic Segmentation Method Based on Generative Adversarial Network. Beijing Jiaotong University.
刘玉洁. 2019. 基于生成对抗网络的图像语义分割方法研究. 北京交通大学.
Lin T Y, Goyal P, Girshick R, et al. 2017. Focal Loss for Dense Object Detection. IEEE Transactions on Pattern Analysis & Machine Intelligence, PP(99):2999-3007.[DOI 10.1109/TPAMI.2018.2858826http://dx.doi.org/10.1109/TPAMI.2018.2858826]
Mattyus G, Luo W, Urtasun R. 2017. DeepRoadMapper: Extracting Road Topology from Aerial Images// 2017 IEEE International Conference on Computer Vision (ICCV). IEEE.
Pavolonis M . 2010. GOES-R Advanced baseline imager (ABI) algorithm theoretical basis document for volcanic ash.
Ronneberger O, Fischer P, Brox T. 2015. U-Net: Convolutional Networks for Biomedical Image Segmentation// International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer International Publishing.
Si G, Fu R D, He C F, Jin W. 2020. Daytime sea fog recognition combined with remote sensing satellites and deep neural networks[J]. journal of Optoelectronics Laser, v.31;No.304(10):78-86.
司光、符冉迪、何彩芬、金炜. 2020. 结合遥感卫星及深度神经网络的白天海雾识别[J]. 光电子·激光, v.31;No.304(10):78-86.
Su J. 2001. Study on circulation dynamic mechanism in China offshore [J] Acta Oceanologica Sinica, 23 (4): 1-16
苏纪兰. 2001. 中国近海的环流动力机制研究[J]. 海洋学报, 23(4):1-16. DOI 10.3321/j.issn:0253-4193.2001.04.001http://dx.doi.org/10.3321/j.issn:0253-4193.2001.04.001
Tian L, Wang X, Yang B Y, Wu X J. 2021. Advances in satellite remote sensing technology for marine atmosphere and environment. Science and Technology Review, 39(15), 13.
田林, 王曦, 杨冰韵, 吴晓京. 2021. 海洋大气和环境卫星遥感技术进展. 科技导报, 39(15), 13.
Tian Y J. Research on daytime sea fog monitoring algorithm based on Fengyun-2 geostationary satellite data. South China University of Technology. (田永杰. 基于风云2号静止卫星数据的白天海雾监测算法研究. 华南理工大学.)[DOI CNKI:CDMD:2.1016.770458]
Wang B H. 1983. Sea fog[M]. Ocean Press Wang Binhua. 1983. Sea fog[M]. Ocean Press
王彬华. 1983. 海雾[M]. 海洋出版社
Wang Z , Simoncelli E P, Bovik A C. 2003. Multiscale structural similarity for image quality assessment// Proc IEEE Asilomar Conference on Signals.
Xiao Y F, Zhang J, Cui T W, Qin P. 2017. Research progress of sea fog satellite remote sensing monitoring. Marine Science, 041(012):146-154.
肖艳芳, 张杰, 崔廷伟, 秦平. 2017. 海雾卫星遥感监测研究进展. 海洋科学, 041(012):146-154. DOI 10.11759/hykx20170523001http://dx.doi.org/10.11759/hykx20170523001
Zhang P, Wu D. 2019. Daytime sea fog detection method based on Himawari-8 data. Journal of Atmospheric and Environmental Optics (03), 211-220.
张培, 吴东. 2019 . 基于Himawari-8数据的日间海雾检测方法. 大气与环境光学学报(03),211-220. DOI 10.3969/j.issn.1673-6141.2019.03.006http://dx.doi.org/10.3969/j.issn.1673-6141.2019.03.006
Zhang C G, Cai Y Y, Zhang J C. 2009. Application of MODIS remote sensing data to sea fog monitoring in the Taiwan Strait of my country. Journal of Applied Meteorology, 20(001): 8-16.
张春桂, 蔡义勇, 张加春. 2009. MODIS遥感数据在我国台湾海峡海雾监测中的应用. 应用气象学报, 20(001):8-16. DOI 10.11898/1001-7313.20090102http://dx.doi.org/10.11898/1001-7313.20090102
文章被引用时，请邮件提醒。
提交
相关作者
暂无数据
相关机构
陕西师范大学 计算机科学学院
北京水保生态工程咨询有限公司
中国矿业大学(北京)
环境保护部卫星环境应用中心
湖北珞珈实验室
- 地址：北京市海淀区北四环西路19号 《遥感学报》编辑部 邮编：100190
- 联系电话：010-58887052 Email：nrsb@aircas.ac.cn
- 技术支持由北京北大方正电子有限公司提供 京ICP备20021838号-9
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies
0
回到顶部