Design And Implementation Of Orchard Irrigation Prediction And Control Algorithm Based On LSTM

According to statistics,China's irrigation water accounts for more than half of the total water resources.With the contradiction between supply and demand of water resources becoming more and more prominent,all countries attach great importance to the development of water-saving society.How to save water resources and realize precision irrigation in agriculture has become an important issue at the present stage.China is a large fruit production and consumption country with a large demand for fruit.According to the survey,the planting area of fruit trees in China has exceeded 300 million mu in 2020.Most of the water absorbed by fruit trees is used for transpiration because of their luxurant foliage and large canopy,and transpiration can indirectly affect the water demand of fruit trees.Taking transpiration as an important influencing factor for the study of water demand of fruit trees is of great significance for saving irrigation water in orchards.In order to realize efficient and intelligent irrigation of orchard,transpiration prediction model was established in this paper.Firstly,Principal Component Analysis(PCA)was used to analyze the collected orchard environmental data,and the key influencing factors were screened out.Then,a prediction model based on Long ShortTerm Memory Networks(LSTM)was established to predict the transpiration of fruit trees.In order to improve the accuracy of prediction,Attention Mechanism is added to the LSTM neural network to form the attention-LSTM prediction model.By comparing the prediction accuracy of the model proposed in this paper with other models,the simulation experiment proves that the prediction model proposed in this paper has high prediction accuracy,simple implementation,and can predict transpiration more accurately under limited environmental data.Secondly,an orchard irrigation control model based on fuzzy control was constructed to make fuzzy control decisions.The predicted transpiration and soil moisture deviation were taken as the input of the model to calculate the precise irrigation duration.Finally,the software platform of orchard irrigation was built,the database and Web page were designed,and the transpiration prediction,fuzzy control and software platform were closely combined to achieve precision orchard irrigation.The intelligent irrigation and traditional irrigation were compared and analyzed in terms of the amount of orchard irrigation,total yield and other aspects,and the intelligent irrigation was superior to the traditional irrigation.It shows that the intelligent irrigation model established in this paper can save water resources and increase fruit yield,which has a certain practical significance.