Study Of High Altitude Meteorological Data Processing Method Based On Data Fusion Technology

Meteorological change is closely related to our life.If we can't master the weather change,human beings will find it difficult to live in the "sky blue planet".Typhoons,floods,sandstorm and so on are related to life of human being.Therefore,it needs convenient and fast methods of meteorological data processing to respond the disasters brought by typhoons,floods and other disasters,and to help humans reduce losses through technology forecasting.The meteorological parameters acquired included temperature,humidity,pressure,latitude,longitude and altitude changing with time in this paper.Data processing method was designed aiming at meteorological data collected by the radiosonde,method of data fusion was researched mainly,and the collected meteorological parameters were analyzed.A set of complete data processing method was designed,which improved the existing collection and processing of meteorological parameters method,and provided more reliable and convenient method for meteorological research in order to predict the weather more accurately.The main contents of the paper are as follows:1.The appropriate sensors of temperature,humidity and atmospheric pressure were chosen according to the measured high altitude meteorological parameters.The temperature transmitting circuit,measurement principle of wet sensitive capacitance and monostable trigger output amplifier circuit,constant current source generating circuit and basic amplifying circuit of silicon piezoresistive pressure sensor were designed to collect the data of high altitude temperature,humidity and pressure.2.Calibration experiments and high altitude experiments were completed and the corresponding data were obtained.Based on the basic principles of the least squares and polynomial fitting,each sensor model was analyzed by MATLAB,regression analysis expressions of the temperature,humidity and pressure were obtained;humidity was compensated through the software and hardware methods,and humidity compensation expressions were obtained;the high altitude meteorological parameters data were processed,and the changes trend and related expressions of meteorological parameters were obtained,and the data of different radiosondes were analyzed contrastively.3.BP neural networks were established to analog compute and simulate the data collected by acquired system.The prediction models of high altitude meteorological data were established through Fletcher-Reeves algorithm,and the influence of hidden node on training was analyzed.The errors of sensors models and high altitude data acquired by different radiosondes were analyzed,the optimal results were obtained through neural network,and the method of data processing was ascertained finally.Results showed that the temperature error was 0.08?,humidity error was 0.8%RH,pressure error was 0.08 kPa,the variance of radiosonde was 7.2159×10-11,these met the requirements of the issue and had certain practical value.