基于面部表情分析技术的葡萄酒中关键香气与饮用舒适度相关性评价  

Evaluation of correlation between key aromas and drinking comfort in wine based on facial expression analysis technology

作　　者：黄翠 武运 薛洁[2,3] 皇甫洁 孙志伟 唐家乐[1] 张成学 李宁 宋涛 HUANG Cui;WU Yun;XUE Jie;HUANGFU Jie;SUN Zhiwei;TANG Jiale;ZHANG Chengxue;LI Ning;SONG Tao(College of Food Science and Pharmacy,Xinjiang Agricultural University,Urumqi 830052,China;China Food Fermentation Industry Research Institute,Beijing 100015,China;National International Joint Research Center for Liquor Quality and Safety,Beijing 100015,China)

机构地区：[1]新疆农业大学食品科学与药学学院,新疆乌鲁木齐830052 [2]中国食品发酵工业研究院,北京100015 [3]国家酒类品质与安全国际联合研究中心,北京100015

出　　处：《食品与发酵工业》2024年第8期146-153,I0002-I0005,共12页Food and Fermentation Industries

基　　金：新疆维吾尔自治区“十四五”重大科技专项(2022A02002-2)。

摘　　要：为明晰新疆地产葡萄酒中影响饮用舒适度的关键香气物质,以新疆四大产区葡萄酒为试材,采用顶空固相微萃取(headspace solid-phase microextraction,HS-SPME)与GC-MS联用技术,结合香气活性值(odor activity value,OAV)确定11款葡萄酒中的关键香气,基于模糊数学传统感官与智能感官面部表情分析对11款酒的感官质量进行综合评价并建立线性回归模型,将关键情绪维度与关键香气物质作相关性分析。研究结果表明,共定性定量了79种香气物质(共有组分43种),酯类数量最多(35种),醇类次之(19种),萜烯类(11种)、醛酮类(7种)、酸类(4种)和其他类(3种)最少。其中吐哈盆地赤霞珠(2号)、焉耆盆地马瑟兰(7号)、焉耆盆地霞多丽(8号)品评得分结果较高;线性回归模型R^(2)=0.831且拟合度良好,说明面部表情分析中其效价及唤醒度能基本反映综合饮用舒适度。通过相关性分析确定与饮用舒适度相关的关键物质为乙酸乙酯、异戊醇、活性戊醇、甲醇、3-羟基-2-丁酮、辛酸等。该研究为提高新疆葡萄酒饮用舒适度奠定了科学理论依据。To clarify the key aroma substances that affect the drinking comfort of local wines in Xinjiang,the key aroma substances in 11 wines were determined by headspace solid-phase microextraction(HS-SPME)combined with gas chromatography-mass spectrometry(GC-MS)and odor activity value(OAV),and the sensory quality of 11 wines was comprehensively evaluated based on fuzzy mathematics,traditional sensory and intelligent sensory facial expression analysis,and a linear regression model was established.The correlation between key emotional dimensions and key aroma substances was analyzed.Results showed that 79 kinds of aroma substances(43 kinds of components)were qualitatively and quantitatively determined,of which esters were the most(35 kinds),followed by alcohols(19 kinds),terpenes(11 kinds),aldehydes and ketones(7 kinds),acids(4 kinds),and others(3 kinds)were the least.Among them,Cabernet Sauvignon(No.2)in Tuha basin,Marselan(No.7)in Yanqi basin and Chardonnay(No.8)in Yanqi basin scored higher.The linear regression model R^(2)=0.831 and the fitting degree was good,which showed that its titer and arousal could basically reflect the comprehensive drinking comfort in facial expression analysis.Through correlation analysis,it was determined that the key substances related to drinking comfort were ethyl acetate,isoamyl alcohol,active amyl alcohol,methanol,3-hydroxy-2-butanone,octanoic acid and so on.This study laid a scientific theoretical basis for improving the drinking comfort of Xinjiang wine.

关 键 词：葡萄酒香气 面部表情综合评价 饮用舒适度 

分 类 号：TS262.6[轻工技术与工程—发酵工程] O657.63[轻工技术与工程—食品科学与工程]