Research and Design of Intelligent Farmland Irrigation System Based on Neural Network

Abstract

Based on the analysis of crop growth cycle and water demand, the factors affecting crop growth water use are divided into three categories: environmental factors, crop factors and soil factors. The training set and test set of the model are selected from the crop irrigation historical data set that meets the expected quality and yield. By designing an intelligent farmland irrigation model based on LSTM neural network algorithm, a method of precise irrigation according to crop growth needs, growth environment and planting soil is proposed. According to the characteristics of factors affecting the water consumption for crop growth, the number of hidden layers of the prediction model is determined, and the network parameters are adjusted; The model is trained on the processed historical irrigation data set to obtain the crop irrigation volume prediction model; The LSTM neural network irrigation prediction model is compared with the traditional RNN neural network irrigation prediction model. The experimental results show that the predicted value and trend of LSTM irrigation prediction model are closer to the real value, with stronger robustness, lower error rate and shorter running time, which can meet the prediction of intelligent farmland irrigation and provide reliable basis for the research of intelligent agriculture.