基于电子舌技术不同超声处理时间的奶酪滋味区分

Discriminating the Taste of Cheese under Different Ultrasonic Treatment Durations by Using Electronic Tongue

为探讨奶酪智能化品质评价的可行性,研究了电子舌专一选择型传感器检测5种传统奶酪和2种进口奶酪过程中的响应信号的特征。采用不同的模式识别方法对奶酪进行分类鉴别,通过主成分分析(principal component analysis,PCA)法、偏最小二乘法将感官评定值与响应值做相关性分析,并探讨了超声处理奶酪对电子舌响应信号的影响。结果表明:应用PCA能够大体反映各类样品的分类信息,PCA降维后的数据能够很好地反映原始数据信息。奶酪的前处理时间不同电子舌传感器的响应值也不同,本实验超声处理条件下对PCA辨别结果影响不大,UMS、GPS和SPS三根传感器与苦、咸味,以PC1为轴呈显著负相关,SWS与甜味以PC2为轴呈正相关。

This paper aimed to study the feasibility of quality assessment of cheeses taste by using electronic tongue. Five traditional cheeses and two import cheeses were tested. Simultaneously, the influence of ultrasonic treatment on sensor response signals was studied. The cheeses were discriminated by different pattern recognition methods. Partial least squares regression(PLSR) and principal component analysis(PCA) were used to establish the relationship between sensor signals and the total taste score value of cheeses. Results showed that PCA could provide general classification information on the various cheese samples and the reduced dimension data provided enough information on the raw data. The ultrasonic processing significantly affected the sensor signals. The signals of sensors named as UMS, GPS, and SPS all correlated negatively with bitterness and saltiness based on principal component 1(PC1), and the response of SWS sensor positively with sweetness based on PC2. The results will provide a basic for the intelligent evaluation of cheese quality.