Study On The Prediction Method Of Soil Moisture In Tea Garden Based On Neural Network

Soil moisture is one of the important factors affecting crop yield and quality.The use of agricultural information technology to carry out soil moisture forecasting is of great significance for the fine management of farmland.This paper,based on neural network technology,studies soil moisture prediction methods in tea gardens,and provides technical support for water-saving irrigation and delicacy management in tea gardens.The neural network model was used to predict the changes of soil moisture in Huangshan tea garden,and to guide the formulation of water-saving irrigation strategies in the agricultural area.The experimental data collected in this paper is the environmental and meteorological information data from December 6,2014 to January 30,2014 in Huangshan Taiping District.Because the tea plantation is located in a hilly region of about 300 meters,the environment is complex and changeable,and the maintenance of data transmission equipment is complex.The collected data is directly used in analysis and modeling.The feasibility is poor,and errors and errors are brought into the data analysis.Experimental results,at this point need to use algorithm to preprocess the source data,make the data more "clean".Firstly,we use the Layda criterion to deal with the larger values,vacant values,and outliers in the data,remove the larger glitches,supplement the vacancies in the data,improve the accuracy of the data,and then use wavelet noise reduction and quadratic exponential smoothing to remove the data.The noise in the data makes the experimental data fall within a reasonable range.After preprocessing,the credibility of the temperature and humidity data is ensured,and the accuracy of soil moisture prediction modeling in tea gardens is improved.There are many factors that affect soil moisture prediction,such as temperature,humidity,p H,light intensity,and soil properties.This article uses correlation analysis to determine the input factors of the predictive model.After research,it is determined that the modeling temperature,humidity and light intensity are the input factors,and the soil moisture is the expected output.In this paper,based on the BP neural network forecasting research on tea garden public opinion,a LM algorithm optimization neural network model is introduced,and the traditional neural network prediction results are compared with the error of the LM algorithm optimized neural network model prediction results to find the optimized neural network.The prediction of soil moisture content prediction is a little higher,the prediction effect is better,and it can more accurately reflect the soil moisture situation in tea gardens.In order to prove that the optimized neural network model has a better prediction effect and is more suitable for tea plantation public opinion forecasting,this paper establishes a soil moisture forecast based on the ARX model and compares the two models' fitness,relative error,The mean square error shows that the optimized neural network can be more suitable for soil moisture prediction and analysis.It has wide adaptability to areas with complex conditions and is of great significance for the promotion of the prediction methods in the later period.