嗅觉可视化技术对啤酒品质的快速检测

为实现啤酒品质的快速无损检测，利用嗅觉可视化技术对青岛啤酒的8 种啤酒（纯生、1903、奥古特、白啤、黑啤、皮尔森、IPA和Strong）进行定性区分和关键质量指标（乙醇体积分数、原麦汁质量分数和双乙酰质量浓度）定量预测。采用4×4的色敏传感器阵列与啤酒的挥发性成分进行反应，并利用化学计量学方法对色敏传感器阵列的差值图像信息进行分析。结果表明，线性判别分析模型能够很好地将8 种啤酒定性区分，校正集和预测集的识别率均达到100%；最小二乘支持向量机模型能够很好地对乙醇体积分数、原麦汁质量分数和双乙酰质量浓度进行定量预测，模型的校正集和预测集校正曲线相关系数均达到0.98以上。因此，嗅觉可视化技术能够实现对啤酒品质的快速无损检测，具有很好的应用潜力。

Rapid Determination of Beer Quality by Using Olfactory Visualization Technology

To rapidly and non-destructively detect beer quality, olfactory visualization technology was used to qualitatively distinguish among eight kinds of Tsingtao beer (namely Premium, 1903, Augerta, Weiss, Stout, Pilsner, IPA and Strong), and quantitatively determine their key quality indicators (namely alcohol content, original gravity and diacetyl concentration). A 4 × 4 colorimetric sensor array (CSA) was used to react with volatile compounds of beer, and the color difference images obtained were analyzed by chemometric methods. The results showed that the linear discriminant analysis (LDA) model was able to qualitatively distinguish among these eight kinds of beer with a 100% recognition rate for both the calibration and prediction sets, while the least square-support vector machine (LS-SVM) model was able to quantitatively predict alcohol, original gravity and diacetyl concentration with correlation coefficients over 0.98 for the calibration and prediction sets. Hence, olfactory visualization technology has good potential for application in the rapid and non-destructive detection of beer quality.