基于径向基函数网络的变量施肥决策研究  

Decision-making Method for Variable-rate Fertilization Based on RBF Neural Network

作　　者：兰维娟[1] 毛鹏军[1] 杜东亮[1] 王俊[1] 

机构地区：[1]河南科技大学车辆与动力工程学院农业工程研究所,河南洛阳471003

出　　处：《安徽农业科学》2007年第21期6505-6507,共3页Journal of Anhui Agricultural Sciences

基　　金：河南省科技攻关计划项目(0620304000);河南科技大学校基金(2004NQ014)

摘　　要：应用RBF神经网络模型训练数据,进行施肥决策,并对神经网络训练的数据及其结果进行分析。结果表明,基于RBF神经网络建立的土壤养分、施肥量与产量之间的关系模型,具有较好的估计结果,可用于指导变量施肥。Artificial Nerve Network is one kind of massive parallel non-linear dynamic system.It has a strong ability in analyzing the complex variable relations.In this paper RBF neural network model was applied to treat data and carry out the decision-making of fertilization.And finally the neural network treating result was analyzed.RBF neural network structure,the inputs of the model were soil nutrient（N,P and K） and application rate of fertilizer（N,P and K）,the output was yield.The model reflected the nonlinear relationship among soil nutrient,application rate of fertilizer and yield. It provided evidenee for decision-making of variable-rote fertilization.

关 键 词：径向基函数网络 施肥量 产量 施肥决策 

分 类 号：S112[农业科学—农业基础科学]