味道
化学
芳香
主成分分析
八醛
线性判别分析
色谱法
气相色谱-质谱法
质谱法
模式识别(心理学)
食品科学
人工智能
己醛
计算机科学
作者
Tong Chen,Haiyu Li,Xinyu Chen,Yong Wang,Qianwei Cheng,Xingpu Qi
摘要
Abstract Fragrant rice has unique aroma characteristics appreciated by many consumers. In order to extract exclusive flavour fingerprints from Guangxi fragrant rice and to further verify its effectiveness by allowing the rapid identification of fragrant rice from two main production regions (Guangxi and Wuchang), gas chromatography‐ion mobility spectrometry (GC‐IMS) technology was applied to analyse flavour components of 53 fragrant rice samples. Based on the two‐dimensional GC‐IMS map, characteristic variables of flavour components from Guangxi fragrant rice were screened via image pretreatment and automatic threshold segmentation algorithm. The extracted variables were further processed by using principal component analysis and quadratic discriminant analysis to establish a non‐linear model for discriminating two types of fragrant rice. The results showed that a total of 46 characteristic variables were extracted from Guangxi fragrant rice, and the first two principal components showed that the two types of rice samples had their own corresponding regions. 2‐Butanone, 2‐heptanone, pinene, 1‐heptanal, 2‐methyl‐1‐butanol acetate and octanal were the key different flavour components. The established model had an excellent recognition rate, which could be used to determine the geographical region of fragrant rice. The proposed method could achieve quick, non‐destructive and accurate analysis of fragrant rice geographical properties. It can be further applied in feature extraction from the two‐dimensional fingerprint map generated by other similar combined instruments.
科研通智能强力驱动
Strongly Powered by AbleSci AI