近红外光谱技术在葡萄及其制品品质检测中的应用研究进展

葡萄具有丰富的营养价值、药用价值和经济价值，是世界上种植面积最大的水果之一。根据人们的消费需求及产品的贮运要求，葡萄被加工成葡萄干、葡萄汁、葡萄酒、葡萄籽油等常见葡萄制品。基于食品质量安全的关注以及高品质果蔬农产品的需求，如何快速有效地评价葡萄及其制品的质量已成为当务之急。随着无损检测技术及装备的快速发展，近红外光谱（NIR）技术因其快速、无损、精确、经济及便于在线分析的优点，也逐渐被应用在果蔬等农产品质量品质检测领域。近年来，国内外学者利用NIR技术在不损坏被检测葡萄及其制品的前提下，结合化学计量法、主成分聚类分析（PCA）、偏最小二乘判别分析（PLS-DA）、主成分回归（PCR）、偏最小二乘回归（PLSR）、支持向量机（SVM）及神经网络（NN）等数据处理方法探究了糖、酒精、酸等一般成分及色素、单宁、芳香物质等特有成分含量与有效光谱信息的关联性，以此建立了葡萄及其制品主要品质指标的定性鉴别与定量分析模型，为便携式近红外葡萄品质检测设备的研制和在线葡萄汁、葡萄酒酿制过程监测系统的开发提供了一定技术支持。文章系统概述了近十年来国内外NIR技术在葡萄、葡萄酒、葡萄汁及葡萄副产物检测中的应用现状，旨在为葡萄及其制品在后续分类鉴定与品质评价研究方面提供参考。研究表明NIR技术通过定量定性分析可实现葡萄复杂理化成分的多组分检测和分类鉴别，在测定葡萄理化特性与内部品质方面的研究已经有了一定的进展，且在葡萄酒、葡萄汁的品质过程监测和定性鉴别方面的研究应用逐渐增多，并逐步应用于葡萄皮中多酚、花青素等葡萄副产物分析，以及葡萄藤、葡萄叶营养生长状况监测等其他方面，进一步证实了NIR技术正成为检测葡萄及其制品品质的一种有效工具，为进一步提高葡萄及其制品品质价值以及实现葡萄果园的实时、高效生产管理提供了技术支持，应用前景极为广泛。如何深入挖掘葡萄及葡萄制品不同类检测数据所反映信息的内在关联性，结合视觉、味觉、嗅觉等多源信息融合技术建立预测精确度更高、更稳健的模型去全面评价葡萄生产、果园管理、成熟期收获及产后加工全过程，实现对葡萄及其制品生产全过程质量控制和在线监测将成为今后的发展趋势。

Recent Advances in Application of Near-Infrared Spectroscopy for Quality Detections of Grapes and Grape Products

Grape, which is one of the fruits with the largest planting area globally, has rich nutritional value, medicinal value and economic value. According to consumers’ consumption demands and storage and transportation requirements for products, grapes were processed into common grape products, such as raisins, grape juice, wine, and grape seed oil. Based on the growing concerns over the quality and safety of foods and the demands for high-quality fruits and vegetables, how to quickly and effectively evaluate the quality of grapes and grape products has become urgent and imperative. With the development of non-destructive testing technology and equipment, near-infrared (NIR) spectroscopy technology has been gradually applied in quality testing of fruits, vegetables and other agricultural products due to its advantages of rapid, non-destructive, accurate, cost-effective, and convenient for online analysis. Nowadays, domestic and foreign scholars have combined the methods of chemometrics and data processing methods, such as principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), principal component regression (PCR), partial least squares regression (PLSR), support vector machine (SVM), and neural network (NN), etc., to determine the relationship between the general components (such as sugar, alcohol, acid, etc.) and specific components (such as pigments, tannins, aromatic substances, etc.) of grapes and grape products and effective spectral information non-destructively using NIR technology. The correlation between the content of quality components and spectral information has been explored to establish qualitative and quantitative analysis models for the main quality indicators of grapes and grape products, which provided some technical supports for the development of portable near-infrared inspection equipment for grape and online monitoring system for grape juice and wine brewing process. This review systematically summarized NIR technology's domestic and foreign application status in grapes, wine, grape juice and grape products in the past ten years for the first time, aiming to provide technical reference for the subsequent classification and identification and quality evaluation of grapes and grape products. Studies have shown that NIR technology could achieve multi-component detection and classification identification of grapes' complex physical and chemical components through quantitative and qualitative analysis. The research on the determination of physical and chemical properties and internal quality of grapes had made great progress, and the research and application of monitoring and qualitative identification of wine and grape juice are gradually increasing. They were gradually applied to the analysis of grape products, such as polyphenols and anthocyanins in grape skins, and the monitoring of the nutritional growth status of grapevines and grape leaves. This further confirmed that NIR technology is emerging as an effective detection tool for the quality evaluation of grape and grape products, improving the quality values of grapes and grape products and providing technical support for real-time and efficient production management have a broad range of applications. For the future research, to sense grape information during the process of growth, harvest, and post-harvest production, and realize quality control and on-line monitoring of grape and its products in the whole production process, studies are heading to investigate the correlation between the spectral information reflected by the detection data of different categories and the inherent quality of grapes and grape products, and build a robust prediction model with high accuracy based on the multi-source information fusion technology of vision, volatile, taste, and smell.