• Laser & Optoelectronics Progress
  • Vol. 62, Issue 7, 0701001 (2025)
Yike Zou1,2,*, Ying Wu1,2, Jingwen Ma1,2, Yuanyuan Huang1,2..., Jinghui Ning1,2 and Qijia Fu1,2|Show fewer author(s)
Author Affiliations
  • 1Key Laboratory of Meteorological Disaster, Ministry of Education, Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu , China
  • 2Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu , China
  • show less
    DOI: 10.3788/LOP242009 Cite this Article Set citation alerts
    Yike Zou, Ying Wu, Jingwen Ma, Yuanyuan Huang, Jinghui Ning, Qijia Fu. Neural Network Inversion Method for Atmospheric Temperature and Relative Humidity Profiles Based on FY-3E/HIRAS[J]. Laser & Optoelectronics Progress, 2025, 62(7): 0701001 Copy Citation Text show less
    Structure of three-layer neural network model
    Fig. 1. Structure of three-layer neural network model
    Variation in temperature inversion error with height on the training set. (a) Clear sky; (b) cloudy sky
    Fig. 2. Variation in temperature inversion error with height on the training set. (a) Clear sky; (b) cloudy sky
    Variation in temperature inversion error with height on the validation set. (a) Clear sky; (b) cloudy sky
    Fig. 3. Variation in temperature inversion error with height on the validation set. (a) Clear sky; (b) cloudy sky
    Variation in RMSE of temperature inversion results with height on the training set. (a) Clear sky; (b) cloudy sky
    Fig. 4. Variation in RMSE of temperature inversion results with height on the training set. (a) Clear sky; (b) cloudy sky
    Variation in RMSE of temperature inversion results with height on the validation set. (a) Clear sky; (b) cloudy sky
    Fig. 5. Variation in RMSE of temperature inversion results with height on the validation set. (a) Clear sky; (b) cloudy sky
    Variation in relative humidity inversion error with height on the training set. (a) Clear sky; (b) cloudy sky
    Fig. 6. Variation in relative humidity inversion error with height on the training set. (a) Clear sky; (b) cloudy sky
    Variation in relative humidity inversion error with height on the validation set. (a) Clear sky; (b) cloudy sky
    Fig. 7. Variation in relative humidity inversion error with height on the validation set. (a) Clear sky; (b) cloudy sky
    Variation in RMSE of relative humidity inversion with height in the training set. (a) Clear sky; (b) cloudy sky
    Fig. 8. Variation in RMSE of relative humidity inversion with height in the training set. (a) Clear sky; (b) cloudy sky
    Variation in RMSE of relative humidity inversion with height in the validation set. (a) Clear sky; (b) cloudy sky
    Fig. 9. Variation in RMSE of relative humidity inversion with height in the validation set. (a) Clear sky; (b) cloudy sky
    ParameterSetting
    Hidden layer transfer functiontansig
    Output layer transfer functionpurelin
    Training Algorithmtrainscg
    Number of nodes in the input layer220
    Number of nodes in the hidden layer300
    Table 1. Parameters of BPNN inversion model
    Yike Zou, Ying Wu, Jingwen Ma, Yuanyuan Huang, Jinghui Ning, Qijia Fu. Neural Network Inversion Method for Atmospheric Temperature and Relative Humidity Profiles Based on FY-3E/HIRAS[J]. Laser & Optoelectronics Progress, 2025, 62(7): 0701001
    Download Citation