Research On Prediction And Prevention Decision Support System For Apple Tree Diseases And Pests

Apple is one of the main economic crops in China. In recent years, with the construction of modern agriculture, apple industry in China has been developing rapidly and our country has become a leading global producer of apple. Apple acreage and yield have topped the world. Apple has been a mainstay of economy in some areas and villages. It helps the farmers supplement their income and improve their life.However, influenced by artificial and natural factors, different kinds of diseases and pests happen frequently which could cause great harm. It not only threatens the yield and quality of the fruit but also bring effect to the economic income and production benefit of the apple growers in different degree. This paper focuses on the diseases and pests of apple trees.Aiming at the demand of apple tree disease and pest control and prediction in our country,take apple canker disease as an example, carry on the research of prediction and prevention decision support system for apple tree diseases and pests. The specific research contents and results are as follows:(1) Research on BP neural network prediction modelUse PCA to process the input data of BP neural network prediction model. Extract the main influence factors to achieve the purpose of dimensionality reduction. This method provides a basic support for improving the precision of the model. Use Sigmoid function as hidden layer activation function and additional momentum method as learning algorithm. The rate of convergence and forecast accuracy of the model are raised. Experiments have shown that the model has fast training rate and high prediction accuracy.(2) Research on wavelet network prediction modelBy comparing the separating wavelet network and integrating wavelet network,integrating wavelet network is selected in the prediction of apple tree diseases and pests. Use PCA to process the input data in order to achieve the purpose of dimensionality reduction.Use Morlet function as hidden layer activation function to build the wavelet network prediction model and predict the epidemic of apple tree canker disease in Taian.Experiments have shown that the prediction accuracy is significant increased compared with BP neural network prediction model.(3) Research on prediction and prevention decision support system for apple tree diseases and pestsFirst, on the basis of requirement analysis, the system is divided into eight function modules which is studied and designed in detail. Second, according to the division of function modules, the database of the system is analyzed and designed. Finally, through the use of S2 SH integration development technology, MySQL database and Tomcat server, the prevention and prediction decision support system for apple tree diseases and pests has been done. The system can provide pathway for users to know the information of diseases and pests, give decision support to disease prediction which is significant for the farmer to prepare prevention and control. It is conductive to make prevention measures according to the seriousness of the diseases in order to reduce economic loss.