Research Of Key Technologies Of Soil Drought Monitoring System Based On Embedded System

In recent years,the extreme weather occur frequently and globally,the drought especially is one of the most influential kind of weather,which has extensively drawn international attention because of the immense financial damage.On the basis of advanced sensor technology and real-time transmission method,and through monitoring data for drought prediction,the technology of drought monitoring can make the early warning to the government departments and government departments can reasonably distribute the drought resistant materials according to warning information,it also can reduce the economic loss caused by drought,improve the efficiency of drought relief,so the research of this technology has both theoretical value and practical significance.This paper mainly studies the key technologies of drought monitoring system based on embedded system,which specifically is the development of the soil moisture sensor and the research of drought prediction model.The new soil moisture sensor is developed by the theory of frequency domain reflection method,and adopts the sensor probe structure of four equal long needles,considering the rationality of the number and distribution of probe and according to the calculation method of probe impedance.The soil moisture sensors takes high-frequency chirp signal to obtain resonant frequency in the media inside the soil,then,get the soil water content.According to the influence of the dielectric constant in the soil under different frequency,the soil moisture sensors adopt 100 MHZ to 150 MHZ frequency chirp signal as a detection signal source.After completing monitoring system,the drought prediction model based on the theory of BP neural network is established,realizing the drought prediction model in central station through the data acquisition and processing of the rainfall,the evaporation,the temperature and soil moisture content,and determining the parameters such as the input and output variables,the network layer,the number of hidden layer and the activation function.In the experiments,the different kinds of indexes on sensor first of all were tested,and the results show that the precision of the soil moisture sensor meets the national standards totally,the accuracy and the instantaneity of the acquisition meet requirements,the range of signal source frequency is appropriate,the appearance design of probe meets the theoretical demands of electromagnetic wave in the transmission line;In addition,the BP neural network model is trained and the real-time collected data is tested,the experimental results show that the model can predict drought accurately,and the neural network model has good fault tolerance.