作　　者： ; ; (赵国军）;

机构地区： 浙江大学生物系统工程与食品科学学院

出　　处： 《传感技术学报》 2006年第4期957-962,共6页

摘　　要： 目前电子鼻在食品领域已有广泛的应用,由于气味组成成分复杂,仅用单个气体传感器是无法评定气体质量的,因此气体传感器阵列应运而生。在对不同的食品进行检测时,电子鼻的传感器阵列的构成是不同的。运用PEN2便携式电子鼻对掺假牛奶进行检测,用方差分析、主成分分析和聚类分析对信号进行分析,得到由8个传感器组成的用于检测掺假牛奶的最佳传感器阵列。用这8个传感器组成的传感器阵列对掺假牛奶进行检测区分,结果用LDA方法(线性判别式分析)进行分析,在LDA分析图上,不同比例掺假的牛奶能被明显地区分。新的传感器阵列具有较少的传感器和与原传感器阵列相同的检测区分能力。 The electronic nose was widely used in the food area. A single couldn＇t assess the quality of odor because its complexity. The sensors array could resolve this problem. The sensors array was different when different food was identified. The adulterate milk were identified by electronic nose, and the signals were analyzed by variance analysis, PCA （principal component analysis） and clustering analysis. The best sensors array made up of 8 sensors was obtained. The new sensors array was applied to identify the adulterate milk. The data was analyzed by LDA analysis （Linear Discriminant Analysis）. The plot of 8-sensor array could distinguish the milk adulterated with powdered milk of different proportion. So the electronic nose with 8-sensor array had the ability of distinguishing the adulterate milk well as electronic nose with the 10-sensor array.

关 键 词： 电子鼻 传感器阵列 主成分分析

领　　域： [自动化与计算机技术] [自动化与计算机技术]