Paper Infomation
Sea Surface Wind Field Data Fusion Based on Full Supervision
Full Text(PDF, 64KB)
Author: Wenlong Cui, 彦鑫 姚
Abstract: Sea surface wind field data fusion can realize the complementary advantages of multi-source wind field data. This paper firstly introduces the traditional data fusion method and the data fusion method based on deep learning. The traditional data fusion method is unable to deeply extract data features to form high level features with strong discrimination. This paper proposes a data fusion method based on U-Net network, which uses GRAPES_1KM model data, HRCLDAS data, satellite wind field data and buoy ship data. When the buoy data is used as the true value of training, the model is difficult to converge due to the sparsity of the buoy data. Therefore, two methods of constructing false tags are proposed in this paper to achieve the convergence of the network model by conducting full supervision training based on U-net network. Experiments show that the loss curve of the network model can converge quickly when the constructed false tags are used for full supervision training, and the effect on data fusion is also obvious.
Keywords: Sea Surface Wind Field, Data Fusion, U-net, Full Supervision
References:
[1] Xinlei Zhang. Study on North Pacific Sea Surface Wind Field based on Multi-Source Remote Sensing Data [D]. Liaoning Normal University,2019.
[2] Jing Liu. Research on Fusion of Sea Surface Wind Field Data in Offshore China Based on Optimal Interpolation Method [D]. National Marine Environmental Prediction Center,2018.
[3] Dongxiang Zhang. Research on Testing and Fusion of Multi-source Satellite Sea Surface Wind Field Products [D]. National University of Defense Technology,2018.
[4] He Fang. Research on Key Techniques of Sea Surface Wind Velocity Inversion in Spaceborne C-band Quad Polarimetric SAR [D]. Nanjing University of Information Science and Technology,2019.
[5] Hao Gao, Shihao Tang, Xiuzhen Han. Development and Application of Fengyun Meteorological Satellite [J]. Science and Technology Review, 201,39(15):9-22.
[6] Jing Liu, Xiaojiang Song, Zhanggui Wang. Sea Surface Wind Fusion Technology Review [J]. Marine Forecast,2018,35(03):81-87.
[7] Zheng Ling, Guihua Wang, Dake Chen, et al. Sea Breeze Convergence in China [C]// National Oceanographic Information. Proceedings of China "Digital Ocean" Forum. Beijing: National Marine Information Center,2006:90-94.
[8] Zhang H M, Reynolds R W, Smith T M. Adequacy of The in Situ Observing System in The Satellite Era for Climate SST[J]. Journal of Atmospheric and Oceanic Technology, 2006,23(1):107-120.
[9] Zhang H M, Reynolds R W, Bates J J. Blended and Gridded High Resolution Global Sea Surface Wind Speed and Climatology from Multiple Satellites:1987-present[C]. Proceedings of the 14th Conference on Satellite Meteorology and Oceanography. Atlant:American Meteorological Society,2006:2-23.
[10] Yan Q S, Zhang J, Meng J M, et al. Use of An Optimum Interpolation Method to Construct a High-Resolution Global Ocean Surface Vector Wind Dataset from Active Scatter Meters and Passive Radiometers[J]. International Journal of Remote Sensing, 2017, 38(20): 5569-5591.
[11] Chuanqi Cheng. Research on Data Fusion Method Based on Deep Learning [D]. Lanzhou University of Technology,2021.
[12] Ruixue Duan, Kaiyue Ma, Yangsen Zhang. Pedestrian Detection Algorithm Based on Multi-Source Data Fusion [J]. Journal of Information Science and Technology University of Beijing (Natural Science Edition),2021,36(01):57-62.
[13] Weijie Shi, Jingjing Huang, Maofa Wang. Comparison of Steel Image Defect Detection Methods Based on Two U-Shaped Networks [J]. Journal of Information Science and Technology University of Beijing (Natural Science Edition),2021,36(01):63-68
[14] Liu Y, Zhang C, Cheng J, et al. A Multi-Scale Data Fusion Framework for Bone Age Assessment with Convolutional Neural Networks[J]. Computers in Biology and Medicine, 2019 108(5):161-173
[15] Zhai J, Dong G, Chen F, et al. A Deep Learning Fusion Recognition Method Based on SAR Image Data[J]. Procedia Computer Science, 2019, 147:533-541
[16] Chen H, Hu N, Cheng Z, et al. A Deep Convolutional Neural Network Based Fusion Method of Two- Direction Vibration Signal Data for Health State Identification of Planetary Gearboxes[J]. Measurement, 2019, 146(11):268-278.
[17] Wu J, Hu K, Cheng Y, et al. Data-Driven Remaining Useful Life Prediction Via Multiple Sensor Signals and Deep Long Short-Term Memory Neural Network[J]. ISA Transactions, 2019, 97:241-250.