摘要
目前通过油料常规质量指标和成分结构信息进行油料种类识别的方法因所需仪器设备多,分析测试过程复杂而缺乏实用性和推广价值;在分析油品理化性能指标与其类别间的相关关系及神经网络的特点后,以最简单方式提取尽可能多的特征参数为原则,通过表观特征参数的途径,设计了一种简单小巧的装置,可同时提取油料密度、粘度、吸光度、电导率和介电常数等参数的特征向量,提出了用RBF神经网络进行油料种类识别的方法,并给出了实现算法;实验结果及应用情况表明,该方法识别效果比较理想,为当前油料种类识别问题给出了一种新的解决途径。
At present,the species of oil can nearly only be identified by its conventional quality indexes and composition.But this needs many instruments,and the recognition process is complicated,so it is almost impossible to put these methods into promoted application.After analyzing the relationship between oil species and its physicochemical properties and the characteristic of neural networks,a compact device for oil＇s characteristic parameters extraction is designed.Through apparent parameters,this device can extract oil＇s density,viscosity,absorbency,conductivity and permittivity with the simplest structure.This paper gives an oil species recognition method using RBF neural network as well as its algorithm.Experiments and applications indicate that this method has ideal recognition effect,and it is possible to identify the species,grade and brand of other kinds of oil products.This method has given a new solution for oil species recognition.
出处
《计算机测量与控制》
CSCD
北大核心
2012年第2期484-486,493,共4页
Computer Measurement &Control
基金
四川理工学院科技项目(2009XJKYL005
2009xjkyL013)资助
关键词
油料种类识别
径向基函数神经网络(RBFNN)
特征参数提取
表观特征参数
oil species recognition
radial-basis-function neural network
characteristic parameter extraction
apparent characteristic parameters
