基于电子鼻的牛奶品质预测方法研究  

Study of Milk Quality Predictive Method Based on Electronic Nose

作　　者：郑海霞[1] 尹芳缘[1] 王敏敏[1] 王绿野[1] 姜燕[1] 沈凤[1] 惠国华[1] 

机构地区：[1]浙江工商大学食品与生物工程学院,杭州310035

出　　处：《中国食品学报》2013年第7期189-194,共6页Journal of Chinese Institute Of Food Science and Technology

基　　金：国家自然科学基金项目(81000645);浙江省公益技术应用研究项目(2011C21051);浙江省自然科学基金项目(Y1100150);国家级大学生创新创业训练计划项目(2012-11);浙江省大学生科技创新活动计划项目(2012R408041)

摘　　要：采用电子鼻系统试验检测掺入不同比例奶粉奶的牛奶样品,用信噪比谱分析方法提取牛奶样品气味特征,不同种牛奶样品得到较好的区分。信噪比谱特征值与奶粉奶掺杂比例之间呈显著负线性相关。采用信噪比特征值线性回归方程作为牛奶品质预测模型,检测不同牛奶样品的信噪比谱特征值,并根据预测模型计算出掺杂比例。该模型可以准确预测掺杂比例,且预测误差小于10%。In this paper, milk samples aduherated with different proportion of powdered milk are discriminated by self-developed electronic nose system. Odor features of milk samples are extracted by signal-to-noise ratio （SNR） spectrum analysis method. Milk samples can be discriminated successfully. The SNR features present negative linear relation-ship with powered milk proportion. The linear regression equation of the SNR feature values is used as milk quality predictive model. SNR feature values are measured through validation experiments, and adulterated proportion values are calculated using the developed model. Results show that this model predicts the adulterated proportion of powdered milk with high accuracy. The predicting error is less than 10%.

关 键 词：电子鼻 牛奶品质 随机共振 信噪比谱 

分 类 号：TS252.7[轻工技术与工程—农产品加工及贮藏工程]