| Agriculture informationization is the important part of agriculture modernization. IT, especially artificial intelligence application in agriculture is the key to agriculture modernization. The application of intelligent system alters the blindness and the subjectivity of decision-maker in agricultural production, reducing the implicit, alleviating the shortage of agricultural technician; it is of great importance for improving production rate and using production knowledge.However, the traditional Expert Systems only treat with the dominant and superficial knowledge, they have lots of demerit such as weak reasoning ability,poor intelligence and the difficulty in getting knowledge and so on. Therefore, this paper introduces the technology of the Artificial Neural Network (ANN) to overcome the shortage of Expert system, and tries to solve the problem, which has complicated relationship, fuzzy boundary, questionably factor and expresses hard in regulation and mathematical model.This paper mainly introduces the usage of Artificial Neural Network in agriculture. The simultaneousness, tolerance, and fuzzy inference of ANN make it become the main trend that Integrating with the Expert System to construct the Intelligent Disease Diagnosis system.Firstly, we study the feasibility and mode of the integration of these two technologies. Secondly, we probe into the type of Neural Network that is suitable for crop diseases and pests diagnose. Taking the advantage of neural network theory, we present a diagnose method based on neural network module. And then we set up maize diseases and pests diagnose BP network model, through the collection and arrangement of the symptoms of maize diseases and pests. Finally we implement the network training and simulating by MATLAB language. The experiment has proved that after the alternate training by noiseless sample and noisy samples, which consist of the key symptoms this model, enjoys high diagnosis accuracy and an excellent capacity of generalization.
|
PDF Full Text Request