Martínez Gila Diego Manuel, Cano Marchal Pablo, Gómez Ortega Juan, Gámez García Javier
Robotics, Automation and Computer Vision Group, University of Jaen, Campus Las Lagunillas s/n, ES-23071 Jaen, Spain.
Sensors (Basel). 2018 Mar 25;18(4):975. doi: 10.3390/s18040975.
Normally the olive oil quality is assessed by chemical analysis according to international standards. These norms define chemical and organoleptic markers, and depending on the markers, the olive oil can be labelled as lampante, virgin, or extra virgin olive oil (EVOO), the last being an indicator of top quality. The polyphenol content is related to EVOO organoleptic features, and different scientific works have studied the positive influence that these compounds have on human health. The works carried out in this paper are focused on studying relations between the polyphenol content in olive oil samples and its spectral response in the near infrared spectra. In this context, several acquisition parameters have been assessed to optimize the measurement process within the virgin olive oil production process. The best regression model reached a mean error value of 156.14 mg/kg in leave one out cross validation, and the higher regression coefficient was 0.81 through holdout validation.
通常,橄榄油的质量是根据国际标准通过化学分析来评估的。这些标准定义了化学和感官指标,根据这些指标,橄榄油可以被标记为粗榨油、初榨油或特级初榨橄榄油(EVOO),最后一种是高品质的指标。多酚含量与特级初榨橄榄油的感官特性有关,不同的科学研究已经探讨了这些化合物对人体健康的积极影响。本文开展的工作重点是研究橄榄油样品中的多酚含量与其近红外光谱响应之间的关系。在此背景下,评估了几个采集参数,以优化初榨橄榄油生产过程中的测量过程。在留一法交叉验证中,最佳回归模型的平均误差值为156.14mg/kg,在留出法验证中,较高的回归系数为0.81。