Characterization of the most odor-active volatiles of orange wine made from a Turkish cv. Kozan (Citrus sinensis L. Osbeck)

The aroma-active compounds of cv. Turkish Kozan orange wine were analyzed by sensory and instrumental analyses. Liquid-liquid extraction with dichloromethane was used for extraction of volatile components. According to sensory analysis, the aromatic extract obtained by liquid-liquid extraction was representative of orange wine odor. A total of 63 compounds were identified and quantified in orange wine. The results of the gas chromatography-olfactometry analysis showed that 35 odorous compounds were detected by the panelists. Of these, 28 aroma-active compounds were identified. Alcohols followed by terpenes and esters were the most abundant aroma-active compounds of the orange wine. Among these compounds, ethyl butanoate (fruity sweet), 3-methyl-1-pentanol (roasty), linalool (floral citrusy), gamma-butyrolactone (cheesy burnt sugar), 3-(methylthio)-propanol (boiled potato, rubber), geraniol (floral citrusy), and 2-phenylethanol (floral rose) were the most important contributors to the aroma of the orange wine because they were perceived by all eight panelists.     

Aromatic profile and odour-activity value of blood orange juices obtained from Moro and Sanguinello (Citrus sinensis L. Osbeck)

Aroma composition of blood orange juices obtained from Moro and Sanguinello (Citrus sinensis L. Osbeck) was analyzed by sensory and instrumental analyses. Liquid-liquid extraction with dichloromethane was used for extraction of volatile components. According to sensory analysis, the aromatic extract obtained by liquid-liquid extraction was representative of blood orange juice odour. A total of 83 and 78 aroma compounds, including alcohols, esters, terpenes, terpenols, aldehydes, acids, ketones, volatile phenols, and lactones were identified in the Moro and Sanguinello juices, respectively. Of these, 15 volatile components presented odour activity values (OAVs) greater than 1, with dl-limonene, nootkatone and linalool being those with the highest OAVs in both cultivars.     

基于机器视觉的核桃仁动态分级研究

【目的】基于机器视觉技术研究出一种适合新疆核桃仁动态分级处理的方法。【方法】利用实时采集且已经完成图像预处理的样品核桃图像得到核桃仁特征集合,运用mRMR特征选择算法筛选原始特征集并对特征的重要性进行排列,通过对支持向量机、决策树和朴素贝叶斯三种机器学习算法进行模型训练和测试,得出最佳分级方法,设计核桃仁自动追踪方法和动态分级流程,构建的核桃仁自动分级系统。【结果】在使用特征bin19、K₁和bin15训练朴素贝叶斯分类器时,核桃仁的分级正确率达到最大为97.33%,在动态条件下运用构建的核桃仁自动分级系统对150个核桃仁进行分级测试,正确率为81.33%。【结论】基于机器视觉研究出的核桃仁特征提取与分级方法、核桃仁动态分级处理动作方法,可以有效完成对核桃颜色和完整度的分级。