Research On Intelligent Irrigation System Based On PSO-GRU Water Demand Prediction Algorithm For Green Pepper Growth Period

Green pepper is a crop that consumes a lot of water.During the growth process,water deficit,excess water and too much water fluctuation will affect its growth.Accurately predict the average daily water demand of green peppers and irrigate on demand is conducive to large-scale planting and improving agricultural production efficiency.At present,the green pepper crops in my country have the following shortcomings in the process of large-scale planting and management: the prediction accuracy of the average daily water demand of green peppers during growth period is not high;the previous experimental data mainly rely on manual collection and recording,and the information is lagging and easy to cause errors;existing The integration degree of existing irrigation system is not high,which makes it difficult to achieve large-scale and intelligent management.For improving the prediction accuracy of the average daily water requirement during the growing period of green pepper,and realizing the intelligent management and control of the irrigation system,this thesis designs and implements a set of intelligent irrigation system based on Particle Swarm Optimization-Gated Recurrent Unit(PSO-GRU)neural network green pepper growth period water demand prediction algorithm.This thesis has done the following researches:(1)According to the time series characteristics of green peppers water demand data,a PSO-GRU model for predicting the average daily water demand of green peppers during the growing period is proposed.Taking the experimental data of water demand and meteorological environment of green pepper from 2014 to 2018 as the data source,taking the environmental data such as average daily temperature,air pressure and wind speed as the feature set,and water demand as the label,using GRU neural network as the training model of water demand prediction.In view of the problem that manual selection of GRU hyperparameters is easy to cause the prediction results to fall into local optimum,particle swarm optimization algorithm is used to optimize the hyperparameters of the GRU model,and the average daily water demand of green pepper was predicted by simulation experiment,and compared with RNN,LSTM,GRU and other models.The simulation results show that the prediction accuracy and fitting effect of PSO-GRU model have been significantly improved.(2)Aiming at the shortcomings of manual data collection and the low integration of the system,a set of agricultural irrigation system was designed and implemented based on the Internet of Things,cloud computing and other technologies.Various agricultural production environment sensors were used to collect various environmental parameters during the growth of green pepper in real time,and the data was transmitted to the cloud information management platform for storage and management through the independently built LoRa wireless network,Realize real-time monitoring of multiple planting areas at the same time.The system supports remote and automatic control of agricultural field peripherals.PSO-GRU is embedded into the information management platform of the irrigation system.The real-time collected crop growth environment parameters are used as the input of the water demand prediction algorithm to predict the water demand of green pepper.At the same time,the water demand prediction results can provide decision-making basis for the formulation of irrigation plans,so as to remotely control various peripheral equipment,and further explore the application of the water demand prediction method in the agricultural irrigation system.After a period of testing and running,the results show that the system is stable and reliable,the functional design is reasonable,and can meet the needs of large-scale planting production.