云南省水稻白叶枯病BP神经网络测报技术研究  

Studies on the BP Neural Network Forecasting Technology of Rice Bacterial Blight in Yunnan Province

作　　者：王园媛[1] 李晓菲[1] 陈涛[1] 刘振华[1] 王云月[1] 

机构地区：[1]云南农业大学教育部生物多样性与病害防控重点实验室,云南昆明650201

出　　处：《云南农业大学学报》2013年第2期257-263,共7页Journal of Yunnan Agricultural University

基　　金：云南省现代农业水稻产业技术体系项目

摘　　要：人工神经网络对复杂非线性问题映射能力强,能提高预测的准确度,为水稻白叶枯病害的防治工作提供指导。因此,本研究基于水稻白叶枯病害发生、危害与温度、湿度、降雨等气象因素相关的特点,利用人工神经网络建立云南省勐海县和石屏县水稻白叶枯病害BP神经网络预测模型,预测病害的发生程度。经实例验证,BP神经网络预测模型预测准确度达到80%以上,较逐步回归模型高。研究表明在勐海、石屏建立水稻白叶枯BP网络预测模型是可行的,并具有较高预测准确度,对防治工作有较高应用价值。The artificial neural network has the strong ability of mapping complex nonlinear problems,which can improve the accuracy of prediction and provide guidance for prevention of bacterial leaf blight of rice diseases.Therefore,based on temperature,humidity,rainfall and other meteorological factors of bacterial leaf blight of rice diseases,artificial neural network was used to establish BP neural Network predictive model of bacterial leaf blight of rice diseases of Menghai and Shiping County,and forecasted the occurrence of the disease.After validating,the prediction model of BP neural network had higher prediction accuracy(>80%),compared with the stepwise regression model.The result showed that,the BP network prediction model of rice bacterial leaf blight is feasible in Menghai and Shiping County,which have higher prediction accuracy and higher value for preventing.

关 键 词：白叶枯病害 BP神经网络 逐步回归 病害预测 

分 类 号：S435.11[农业科学—农业昆虫与害虫防治]