Inversion And Application Study Of Atmospheric Temperature And Humidity Profile Based On Microwave Radiometer

The atmospheric temperature and humidity profile describes the changing relationship between atmospheric temperature,relative humidity,water vapor density and altitude,providing important information in the field of climate research.The common atmospheric detection equipments include sounding balloons,weather radar,sounding aircraft,microwave radiometer and so on.In recent years,because of obvious advantages,microwave radiometer has been widely used in the field of remote sensing atmosphere information.Therefore,in this paper,the methods and application of the atmospheric temperature and humidity profile inversion based on the microwave radiometer are developed.Firstly,with the research of the hardware structure and working principle of the MP-3000 A microwave radiometer,the theory of thermal radiation and the relationship between brightness temperature and brightness is given.Then,the brightness temperature of atmospheric microwave radiation is derived by the theory of radiative transmission and the characteristics of atmospheric microwave absorption and radiation.Secondly,according to the absorption characteristics of water vapor and oxygen,the formula of absorption coefficient of them is derived,and the transmission equation of atmospheric microwave radiation is obtained.After processing the original sounding data,the MonoRTM radiation transmission model is introduced,and the simulated brightness temperature is performed,which is ready for the next study.Thirdly,several common methods of inversion of atmospheric temperature and humidity profile are introduced,including the advantages and disadvantages of those methods.Then,the artificial neural network is selected to be the research method.The working principle,training methods and structural parameters of the neural network are analyzed.A BP neural network model and a BP neural network model with shortcut connection are designed and implemented.Finally,the training data is processed,including elimination of abnormal data,classification of weather conditions and the normalization of sample data.A large number of experiments are designed to determine the parameters of the neural network to obtain the best performance and to compare the two neural network models.On this basis,different neural networks are trained for different weather conditions,which improves the accuracy of inversion operation,and error analysis for trained neural network is conducted.Afterwards,temperature and humidity profile inversion software is designed and implemented,with the function and the interface of the software introduced.