利用BP神经网络描述作物-水模型  

Artificial Neural Network Model For Cropresponse to Water

作　　者：缴锡云[1] 雷志栋[1] 杨诗秀[1] 丛振涛[1] 

机构地区：[1]清华大学水利水电工程系,北京100084

出　　处：《河北工程技术高等专科学校学报》2002年第1期1-5,共5页Journal of Hebei Engineering and Technical College Quarterly

基　　金：国家自然科学基金重点资助项目 (5 983932 0 )

摘　　要：研究了利用 BP神经网络描述作物 -水模型的方法。将各生育阶段的耗水量作为输入 ,将作物单产作为输出 ,构建了具有 1个隐含单元层的 BP神经网络结构。利用北京永乐店试验站 1998～ 1999年度的冬小麦灌溉试验资料作为样本进行训练和检验 ,得到了作物 -水模型的人工神经网络模型。结果表明 ,作物This paper presents a method of using BP artificial neural network to describe crop water model.With the relative water consumption of each growing stage as input and relative crop yield output,a BP artificial neural network structure containing a single hidden layer is established.The experiment data on winter wheat irrigation that was collected from 1998 to 1999 in experiment station of Yong Le Dian is regarded as samples for training and testing and the artificial neural network model(ANN) for crop water production function is obtained.The study results indicate that the ANN model for crop water production function of the fitting precision and the predicting precision are satisfactory.

关 键 词：作物-水模型 耗水量 产量 人工神经网络 BP神经网络 作物生长过程 水分变化 灌溉制度 

分 类 号：S274[农业科学—农业水土工程]