Research On Farmland Intelligent Irrigation System Based On Crop Water Demand Forecast

China has a large population,and food demand has always been the driving force of agricultural development.In recent years,with the arrival of the development trend of agricultural information technology,China’s agricultural development has come to a new turning point,that is,from traditional agriculture to modern agriculture.In the absence of information technology,farmers rely on their own experience to irrigate crops.This irrigation method will not only result in the ineffective use of water resources,but also will not improve the crop yield as expected.The water demand of crops varies from sowing to maturity at each growth stage.Therefore,designing an irrigation prediction model based on the water demand of each stage of crop growth using long-short memory neural network and Penman formula,and building a set of intelligent irrigation system based on crop water demand prediction combined with wireless transmission technology is a useful attempt to achieve water saving,yield increase and intelligent irrigation.The main work of this paper is as follows:(1)The optimization and selection of intelligent irrigation hardware equipment.Through consulting the literature of intelligent irrigation technology and equipment research and development at home and abroad,the typical demonstration area of intelligent irrigation in China is investigated,various environmental factors of farmland are analyzed,and a set of intelligent irrigation equipment scheme is formed,including field micro meteorological station,soil parameter sensor,intelligent electromagnetic valve,crop growth and long-term monitoring instrument.(2)The construction of forecasting model of crop water demand.Taking summer corn as an example,the water demand characteristics of crops in each growth stage were analyzed.Based on the calculation results of crop water demand by Penman formula,the long-term memory network(LSTM)was used to predict crop water demand.The main factors affecting crop water demand are found out by grey correlation algorithm,which are the temperature(including the maximum daily temperature and the lowest daily temperature),the humidity in the farmland,the wind speed in the farmland and the sunshine hours in the farmland.Then,the numerical sequence of these climatic factors is selected from the historical data,which constitutes the training set and test set of the prediction network.The results show that the LSTM network can be used to forecast water demand more accurately than the empirical models of Hargreaves,makkink and Priestley Taylor,and can be used for intelligent decision-making in irrigation system.(3)Design of intelligent irrigation system.The microprocessor module,wireless communication module,information acquisition module,control module and intelligent algorithm are combined to realize real-time collection of farmland data and accurate irrigation of crops.Design mobile app,realize the user to the farmland data query and irrigation remote control.The intelligent irrigation system based on crop water demand prediction has a certain application value in water saving and yield increase.