现代电子技术2016，Vol.39Issue(14)：9-11,3.DOI:10.16652/j.issn.1004-373x.2016.14.003
基于神经网络的水汽/液水含量反演方法研究
Water vapor/liquid content inversion method based on neural network
摘要
Abstract
The brightness temperature received by ground⁃based microwave radiometer has no perfect linear relationship with the atmospheric water vapor and cloud liquid water contents for describing weather and climate. The nonlinear problem can be solved with the neural network algorithm. The historical radiosonde data of Zhengzhou district is used to simulate 24 GHz and 35 GHz dual channel brightness temperature,ground temperature,barometric pressure and relative humidity to constitute the in⁃put vector. The atmospheric water vapor and the cloud liquid water contents calculated with same radiosonde data are taken as the output vector to train the BP neural network,and then the validation sample is input into the trained network to carry out simulation. The comparison and detection results show that it has a good correlation with the atmospheric water vapor and cloud liquid water total contents calculated as the true values. The correlation coefficients are 0.953 82 and 0.934 75. The validity of the method was testified.关键词
地基微波辐射计/BP神经网络/大气中水汽含量反演/云中液态水含量反演Key words
ground-based microwave radiometer/BP neural network/inversion of water vapor content in air/inversion of liquid water content in cloud分类
电子信息工程引用本文复制引用
王旭,牛海斌,杨桂玲..基于神经网络的水汽/液水含量反演方法研究[J].现代电子技术,2016,39(14):9-11,3.基金项目
国家自然科学基金 （）
