摘要
在对蔬菜病虫害症状的特征描述的基础上 ,利用知识工程的方法和技术 ,结合神经网络、不确定性推理、多媒体、友好界面和协作冲突消解技术 ,实现了基于神经网络的蔬菜病虫害诊断专家系统框架。在本框架的神经网络中 ,采用了渐近学习、粗集消除冗余样本等技术。系统根据领域专家提供的事实数据 ,作为系统所构造的神经网络的样本数据来训练网络。为了使用户输入病虫害症状特征更形象、直观 ,以选择方式为主的输入方法 ,在选择按钮旁边标出提示样板或实物示例 ,辅助用户输入。利用此框架 ,编制出了番茄病虫害诊断专家系统。
On the basis of the character description of symptom of vegetable diseases and insect pests, by using methods and technology of knowledge engineering, with the help of the technology of Neural network, uncertainty reasoning, Multi media, friendly human computer interface, and cooperative conflicts resolution, the diagnose expert system framework for vegetable diseases and insect pests is implemented. The gradually learning and removing redundant samples with Rough sets theory is used in this Neural Network. The neural network was trained with the fact provided by the field expert in order to input symptom of vegetable diseases and insect pests visually.The way of inputting is choosing method. Prompting templets or object demonstration is lined out by the selecting button to help users to input conveniently. Using this framework, the diagnose Expert System of tomato diseases and insect pests was worked out.
出处
《重庆大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2001年第4期42-44,共3页
Journal of Chongqing University
基金
国家 8 63高技术研究发展计划资助项目 ( 863 3 0 6 ZD0 5 0 3 F)
