To establish a method for rapid detection of honey adulteration,in this paper, Acacia honey, Jing nectar, Date honey of market liquidity greatly and rape honey, rice syrup of typical adulterated substances were studied, the electronic nose (e-nose) information of natural honey and adulterated honey that was made by incorporation 10%, 20%, 30%, 40%, 50%, 60%, 70%rape honey and rice syrup were collected by FOX4000 e-nose. The data processing method of principal component analysis (PCA) was employed to study information variation characteristics of the electronic nose of natural honey and adulterated honey, and the minimum amount added that adulterated substances cause honey aroma system changes. On the basis of information variation characteristics of the electronic nose was as the guidance, the pattern recognition algorithm of Linear Discriminant Analysis (LDA) was employed to study the ability of qualitative recognition of electronic nose for adulterated honey. The results showed that there was a linear relationship between e-nose signals and the level of adulteration. Adulterated honey samples array from right to left with the increase in the content of adulteration in the direction of the principal component 1, adulterated samples of incorporation rape honey and rice syrup both located on the left side of pure acacia honey and arrayed from right to left in the direction of the principal component 1. The linear relationship of adulterated information was not affected by nectar, and the adulterated samples of date honey, Jing nectar, acacia honey had the same regular pattern, and the linear relationship of adulterated information had universality in honey. The minimum amount to be added of rape honey and rice syrup as adulterated substances that can cause honey aroma system changes were 2% and 1%, respectively. Honey aroma system can easily be changed by added substances. This showed that the electronic nose had a strong discriminable ability for honey adulteration. The linear discriminant analysis linear that was pattern recognition algorithm was employed to distinguish honey adulteration. The results showed that pure honey and adulterated honey can be distinguished by LDA pattern recognition algorithms, and the accuracy of discriminant model was 94.7%, and this achieved the rapid identification for honey adulteration. The pure honey and adulterated honey can be well distinguished, and was far away each other, this show the difference between pure honey and adulterated honey can be discovered by Linear Discriminant Analysis (LDA). The method of electronic nose combined with LDA pattern recognition algorithm provided a reference for the identification of fast and accurate of honey adulteration. The study may provide an identification method of effective and rapid for honey adulteration, and had great significance to ensure the healthy development of honey industry.%为建立蜂蜜掺假快速检测方法，该文利用电子鼻并结合主成分分析（principal component analysis，PCA）数据处理方法研究了掺入10%、20%、30%、40%、50%、60%、70%油菜蜜和大米糖浆的掺假蜂蜜的电子鼻信息变化特征，并以掺假蜂蜜的电子鼻信息变化特征为指导，结合判别因子分析（linear discriminant analysis，LDA）模式识别算法研究了电子鼻对掺假蜂蜜的定性识别分析能力。结果表明，掺假蜂蜜的电子鼻信息呈现线性变化，并且电子鼻对掺假蜂蜜有较强的敏感力。LDA 模式识别算法可以将纯蜂蜜样品与掺假蜂蜜样品很好的区分开， LDA掺假判别模型正确识别率为94.7%，该技术可以为蜂蜜掺假鉴别提供技术支撑。
展开▼
