Research And Application Of Meteorological Data Based On Matrix Completion Technology

With the development of the society,the impact of natural disasters to human living environment is increasing.The increasingly extreme weather events,brought a lot of trouble to human life.Meteorological monitoring is more and more important.However,restricted by the level of science and technology,the accuracy and effectiveness of weather events remains to be improved.In the field of meteorological research and application,observation data is an important means for monitoring weather systems,it is the data base of climate analysis and the standard of validation of weather forecast.Real-time and accurate monitoring of weather system data can provide strong technical support for disaster prevention and reduction of the decision-making department.In recent years,the rapid development of wireless sensor acquisition equipment has brought new opportunities for the meteorological observation system,it greatly make up the shortage of traditional manual observ ation,such as low frequency,large data errors.And one can achieve all-weather automatic real-time observation.However,in the field of meteorology,during the acquisition of meteorological data,there may be exist missing data in the part of acquisitio n data or certain error with the actual weather conditions,due to the complex of observation environment,the failure of acquisition equipment and the network fault.In the model of meteorological forecast data,the error of the meteorological data can le ad to the conclusion of the weather forecast.How to recover the missing data and process the errors in meteorological data becomes a difficult problem.Matrix completion is mainly to solve a problem about whether one can recover the entire matrix elements just from a part of known matrix elements.Namely,it is a problem to fill the missing data by only observing a small part of the data matrix.Meteorological data has its own advantage on spatio-temporal correlation.Using the matrix completion technique to predict the missing meteorological data,to mine and process the meteorological data and then to carry on deeper analysis,which provides help for the basic service and decision service of meteorological,and has important theoretical and practical value.We build model for meteorological acquisition system,collect and store the information of monitoring meteorological.The meteorological data of the regional automatic weather station in Zhuzhou is collected as an example.We propose a matrix completion algorithm to solve the problem of missing data recovery from a small amount of meteorological element data.Some common matrix completion algorithm has been introduced in this paper,mainly a meteorological data matrix completion problem based on the SVT algorithm is proposed.its performance verified by experimental tests.The experimental results show that when the collection of meteorological data meet the requirements of recovery,our approach can recover the missing meteorological data with high accuracy,which proved the practical utility of this method.In this paper,based on the technology of wireless sensor network and the actual demand of the automatic meteorological monitoring system,we give the overall design scheme of acquisition and processing system of meteorological data.We build model for meteorological collection system,and we have carried detailed analysis and design on the processing flow of the system’s data,the model of software development and the database.In our system,to construct the network of meteorological data gathering,there are 196 collection terminal of meteorological monitoring distributed in zhuzhou.These collection terminal communicate with monitoring center through the SIM card of mobile phone with the GPRS communication,so that achieve the GPRS wireless network,Based on GPRS,we build a collection and communication system of meteorological data.Which has the ability to achieve real-time acquisition and the transmission of the meteorological element data in the current area.And achieve the wireless,automation and intellectualization for real-time collection and processing of meteorological data.