Research On Soil Moisture Early Warning System Based On Big Data Analysis

To solve the problem of farmland soil moisture forecast and early warning,to better guide the agricultural production,this research through the precipitation,air temperature,humidity,wind speed,sunshine time meteorological data analysis,such as high efficiency of farmland soil moisture monitoring system as the goal,set up the model of entropy method,according to the size of the index variability to determine the objective weights,if a index,the smaller the information entropy indicates that index variation,the greater the degree of information,the more can play a role in the comprehensive evaluation is,the greater the its weight.And vice versa.This model is used to construct a model analysis system for soil moisture data monitoring and a large data analysis framework for intelligent irrigation information system.Using Spark as the computing engine for large-scale data processing,first of all,the regression analysis was used to study the relationship between meteorological data random variables?X?and soil moisture?Y?.Secondly,the interactive summary table composed of qualitative variables is used to show the differences between different categories,showing different variables?X₁,X₂,...,X_k?,etc.The corresponding relation between each category,form list,each element in the list of rows and columns,respectively,the ratio between the structure in the form of point correspondence analysis in low dimensional space,it is concluded that farmland soil moisture monitoring method for the optimization design of meteorological information service.System implementation using the Spark of Streaming current distributed computing,to collect data through Filebeat data collected directly to Kafka,after the acquisition and use Spark Streaming data in a read Kafka data directly into Redis and database,by Spring Boot to provide services,read the data in the database,using Spark show Web visual interface.In this paper,through a period of time,in a certain area of meteorological data of soil moisture forecast,monitoring effect,such as research,through the long time of observation data,the establishment of local rainfall and other meteorological factors and the quantitative relation between soil moisture content,achieve according to the meteorological factors to predict soil moisture and soil moisture monitoring,real-time diagnosis and forecast service purposes.The results showed that:?1?the time period of historical monitoring day and the daily soil moisture content diagnosis can be realized and universally applicable in the establishment of soil moisture prediction model based on rainfall data of meteorological stations.?2?to improve the accuracy of the model,there should be at least four conditions,one is the maximum of three independent variables and is a relatively independent variable;Second,abide by the law of conservation of mass and statistical methods;The third is the method to deal with the number of days without fixed monitoring;Fourth,it is better not to determine parameters in the process of data processing.?3?when other conditions are unchanged,the average annual precipitation is less,and the corresponding model is more accurate.A scientific and practical soil moisture content model based on the precipitation of meteorological stations provides a theoretical basis and model method for the establishment of soil moisture monitoring network and the realization of real-time dynamic soil moisture prediction.Conclusion:Soil moisture warning method based on big data analysis can be extended to other agricultural forecast and warning.