Study And Application Of Corn Fertilization Model Based On Fuzzy Neural Network Optimization Algorithm

The rapid economic and social development has led to a series of increasingly serious problems,such as the increase of population and decrease of land,resource shortage and ecological environment deterioration.Modern information technology has provided unprecedented new driving force for China’s agricultural modernization.The deep integration of information technology and agriculture,represented by precision agriculture and smart agriculture,will certainly provide a new way to solve the problem of resource shortage and environmental pollution.Precise fertilization is the core link of precision agriculture.It is necessary to understand the relationship between soil nutrients and yield and carry out precise fertilization to improve crop yield.Corn,the top one crop in China also in the world,plays an important role in ensuring national food security.Among the existing agricultural techniques related to precise fertilization,there is no good fitting for the complex non-linear relationship between soil fertility parameters and yield in the process of fertilization,which affects the precision of fertilization model to some extent and limits its application scope.Radial Basis Function(RBF)technology can predict the non-linear relationship more efficiently and accurately.Therefore,based on the project of "Integration and Demonstration of Corn Precision Operation Technology Based on Internet of Things" of National Spark Program,this paper chose soil nutrients and maize yield as the research object,carried out optimization algorithm research on precision fertilization model based on Fuzzy c-means(FCM)and RBF,and provides decision-making basis for precision fertilization of maize.The main research is as follows:(1)collect and sort out the data about corn precise fertilization from operation agricultural security demonstration area mentioned in the national "863" project and spark plan.This paper summarized and analyzed the deficiencies and shortcomings of previous corn precise fertilization models,and clarified the research idea of the corn precise fertilization model integrated by fuzzy clustering and RBF neural network.(2)the optimization of RBF neural network algorithm based on fuzzy c-means clustering algorithm is studied.In view of the shortcomings of traditional RBF neural network in processing multidimensional data and excessively relying on the selection of hidden layer data centers,the kernel function-based FCM optimization RBF neural network algorithm was introduced,and the corn yield prediction model was constructed by using this optimization algorithm.The results showed that the optimization algorithm has the advantages of high accuracy and strong robustness,and it is suitable for the construction of complex nonlinear precise fertilization model of corn.(3)an optimization algorithm was proposed for the establishment of corn precise fertilization model based on FCM and RBF neural network technology.Firstly,the traditional BP neural network and RBF neural network,fcm-rbf neural network and fcm-ols-rbf neural network were compared to determine the optimal fusion algorithm of fcm-ols-rbf neural network.Then the yield of maize and the nutrient content of soil were calculated by the fusion optimization algorithm,and the model of precise fertilization for simulated maize was established.Research and experimental results showed that fcm-ols-rbf neural network has the best prediction effect,high accuracy and fast speed in solving the problem of precise fertilizer application of corn.This method is not only superior to the traditional fertilization model,but also has strong robustness and generalization ability,which can provide scientific decision-making basis for precise fertilization of corn.(4)an intelligent corn precise fertilization system based on the optimized neural network model was developed.Combined with the above research results,an intelligent corn precise fertilization system based on optimized neural network model was developed on the platform of.net and the language of C#.The system has realized the functions of user management,yield prediction,precise fertilization,expert consultation and help,and has been applied he demonstration area of Nong’an County under the National 863 Plan and Spark Plan,and has achieved good economic and ecological benefits.