电子鼻技术在鱼粉酸价检测中的应用  

Electronic Nose Technology Applied in the Detection of Acid Value in Fishmeal

作　　者：刘辉[1] 牛智有[1] 

机构地区：[1]华中农业大学工程技术学院,武汉430070

出　　处：《农机化研究》2011年第1期195-198,共4页Journal of Agricultural Mechanization Research

基　　金：国家"十一五"科技支撑计划项目(2006BAD14B03)

摘　　要：酸价是评价鱼粉新鲜度的一个重要指标,探索酸价含量的快速检测对鱼粉品质快速检测具有重要意义。为此,对不同新鲜度的鱼粉样本进行酸价含量的化学测定和电子鼻数据采集,利用BP神经网络和支持向量回归(SVR)建立酸价含量和电子鼻数据两者之间的回归模型,并用预测集样本进行验证。采用BP神经网络和SVR结果相近,其决定系数R2、预测标准差SEP、最大相对误差RE-max、平均相对误差RE-mean分别为0.9 7,0.2 0,7.51%,2.45%和0.97,0.27,10.73%,2.94%。结果表明,利用电子鼻技术检测鱼粉酸价含量是可行的。Acid value is a very important indicator to freshness of fishmeal.Exploring rapid detection of acid value has great significance to fishmeal quality inspection.Fishmeal samples with different freshness were determined acid value by chemical and acquired odor information by electronic nose.BP neural network and support vector regression models between acid value and electronic nose data were created,and they were validated by prediction set.The determination coefficients,standard deviation of prediction,maximum relative error,average relative error respectively were 0.97,0.21,8.05%,2.61 and 0.96,0.23,9.64%,2.58% of BP neural network and support vector regression models.The results showed that it is feasible to detect acid value of fishmeal based on electronic nose.

关 键 词：电子鼻 BP神经网络 支持向量回归 鱼粉 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]