Research And Application Of Agricultural Irrigation Forecast System Based On BP Neural Network

Water is an important natural resource for human survival.At present,the shortage of water resources has become a global problem.China's demand for freshwater resources is greater than others,so the problem is even more severe.The amount of water used in agriculture accounts for a large proportion of total water consumption,and the utilization rate of water used in agriculture is still low.And meanwhile China is facing unprecedented challenges in agricultural water because of the lack of objective reference to water consumption and the waste of water caused by subjective operation.It has great development significance in China's agriculture,society and economic that developing water-saving irrigation to improve the utilization rate of agricultural water.Aiming at the problems of uneven distribution of water,complex environment,insufficient data and lack of objective basis in China's agriculture,a set of irrigation system based on machine learning that predict soil moisture content is designed in this paper.The system consists of two parts: hardware and software.The hardware part is the Internet of things(IoT)farmland parameter acquisition system,including wireless sensor network,IoT gateway,server and database,humancomputer interface,etc.This part is responsible for acquisition of soil moisture content,air temperature and humidity,rainfall,light,rainfall,wind speed and direction,and storing them in a database.The software part is the sever built by neural network system that predict soil moisture content.It uses the data preprocessing algorithm in the R environment to analyze and process the relevant data,and then constructs a neural network model with soil moisture content as the target parameter.In this system,other parameters are taken as input to obtain the target parameter,thus providing reference basis for irrigation water demand.The results of a series of tests of irrigation system and simulation experiments of neural network model show that this system can accurately collect all kinds of environmental parameters of crop growth by using wireless IoT information collection mode,then store these data in the database and show these to users.By constantly building and updating the neural network model,the existing prediction system is improved so as to predict the soil moisture content more accurately.It not only provides objective and scientific decision basis for agricultural irrigation system,but also realizes efficient water saving,waste reduction and automatic irrigation.