摘要
精确估计云顶高度可以在数值天气预报以及气候模型领域中起到举足轻重的作用。云顶高度首先影响着加受油机对接,影响着飞机通讯质量,其次降低对目标的识别,还会破坏电磁干扰与抗干扰。因此,获取更精确的云高信息,不仅节约了时间,还避免了错误估计而造成不必要的损失。本文以训练样本集中日本葵花8静止气象卫星光谱通道数据为输入,以美国CALIPSO卫星的结果为输出,通过构建神经网络模型方法,利用BP神经网络,针对葵花8静止气象卫星,建立葵花8卫星云顶高反演模型,使用检验样本集,对反演模型取得的云顶高反演结果进行检验与分析,研究模型的云顶高反演能力、特性和效果。
Accurate estimation of cloud top height can play a pivotal role in numerical weather prediction and climate modelling.The height of the cloud top first affects the docking of the oil receiving machine,affecting the communication quality of the aircraft,and secondly reducing the recognition of the target,and also destroying electromagnetic interference and anti-interference.Therefore,obtaining more accurate cloud high information not only saves time,but also avoids erroneous estimation and causes unnecessary loss.In this paper,the data of the spectral data of the Japanese sunflower 8 geostationary meteorological satellite in the training sample are taken as input,and the results of the US CALIPSO satellite are taken as the output.By constructing the neural network model method and using the BP neural network,the sunflower 8 satellite cloud top is established for the sunflower 8 geostationary meteorological satellite.The high inversion model uses the test sample set to test and analyze the Genting high inversion results obtained by the inversion model,and to study the model’s cloud top high inversion ability,characteristics and effects.
作者
刘申英紫
孟恒
Liu Shenyingzi;Meng Heng(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)
出处
《科技通报》
2020年第6期67-71,共5页
Bulletin of Science and Technology
关键词
云顶高度
日本葵花卫星
BP神经网络
反演模型
genting height
Japanese sunflower satellite
BP neural network
inversion model
