Remote Diagnosis and Control Expert System for Citrus Agricultural Diseases and Insect Pests Based on BP Neural Network and WebGIS

In this paper, on the basis of full investigation and a lot of comparison with expert systems at home and abroad, introducing Web technology, WebGIS and BP neural network technology into expert system and taking full use of their respective advantages, we have designed and developed a remote diagnosis and control expert system for citrus agricultural diseases and insect pests based on WebGIS and BP neural network. We put emphases on the design of forward inference engine based on neural network and reverse inference engine based on rule system. In the whole expert system, neural network serves as a forward inference engine for the expert system. It accepts standardized raw data input of the diseases and insect pests and gives the diagnosis conclusion after reasoning process. Reverse reasoning is made by rule system, which is going to verify the diagnosis conclusions provided by neural network. In this way, the reasoning speed of entire system and correct diagnosis rate are greatly increased. At the same time, the rule system provides control methods for diseases and insect pests according to the diagnosis conclusions and also provides its geographical distribution information based on GIS. The experimental results show that the expert system has an accurate and reliable performance, so it has a good promotion prospects for applications.