基于BP神经网络粒穗连接力的测定  

Determing Connection Force of Rice Grain Spikes Based on Neural Network

作　　者：杨立[1] 衣淑娟[1] 毛欣[1] 

机构地区：[1]黑龙江八一农垦大学信息技术学院,黑龙江大庆163319

出　　处：《湖北农业科学》2014年第22期5545-5548,共4页Hubei Agricultural Sciences

基　　金：黑龙江省自然科学基金项目(GZ11B110);黑龙江省教育厅项目(12531458);黑龙江省农垦总局开发项目(HNK12KF-20)

摘　　要：利用OM-8650电子拉力试验机测定了北方寒地水稻(Oryza sativa)与机械化收获有关的生物参数,通过试验测得的参数(子粒的长、宽、厚、含水率等)检测水稻的机械特性指标(子粒连接力)建立BP神经网络模型,来确定网络的拓扑结构,并训练建立的BP神经网络。结果表明,通过网络训练可以确定出收获脱粒的最佳时间和脱粒元件。The biological parameters of north cold rice related to mechanical harvesting of OM-8650 were measured by electronic tensile testing machine. Parameters inciuding grain length,width,thickness, moisture content measured by experiment index of mechanical properties test of rice（grain connection force） were used to establish the BP neural network model. The topology of the network was determined. The BP neural network was established and trained. The results showed that the optimal time of harvesting and threshing, and selection of threshing component were determined by training the network.

关 键 词：寒地水稻(Oryza sativa) 连接力 脱粒 含水率 BP神经网络 

分 类 号：S347.1[农业科学—作物栽培与耕作技术]