基于显微高光谱成像技术的抹茶品质快速无损检测机理研究

Matcha is a kind of super fine and green tea powder, as a natural green food, which has a very high value added. However, currently, there is a lack of fast and objective method for the evaluation of matcha quality. Thus, studying a fast and nondestructive evaluation method for matcha, which has important scientific significance and application value. The project will put forward the new idea of intelligent evaluation of matcha quality using hyperspectral microscope technology and will extend the traditional hyperspectral imaging detection technology from the macro field to the micro domain. First of all, this study will explore the relationship between particle size, color and functional components, and the influence of the different quality of matcha on optical absorption and scattering properties at micro scale. And this study will reveal the correlation between particle size, color, functional composition content and micro spectral image information, and clarify response pattern of micro spectral imaging for matcha. Then, spectral image denoising, dimensionality reduction, feature extraction methods will be studied, to effectively extract information from the large spectral image data. Next, the spectral pretreatment, feature data fusion and model establishment methods will be studied, and the intelligent evaluation model for matcha will be finally established using micro spectral imaging information. The rapid and nondestructive detection mechanism for matcha quality using hyperspectral microscope technology will be revealed. This study will provide a theoretical basis for the construction of a new intelligent nondestructive testing method for matcha quality.

抹茶是一种超细微翠绿色茶粉，作为天然绿色食品，具有极高的附加值，但当前缺乏抹茶品质的客观、快速评价方法，研究抹茶品质的快速无损评价手段具有重要的科学意义和应用价值。项目提出抹茶品质的显微高光谱成像快速无损检测新思路，将传统高光谱成像检测技术从宏观领域拓展到微观领域。首先，研究抹茶粒度、色泽、功能成分等品质属性的相关性，探索抹茶品质特征对显微尺度光传输吸收和散射特性的影响规律，揭示抹茶品质特征与显微光谱图像信息的相关性，解析抹茶品质的显微光谱成像响应规律；然后，研究显微光谱图像的滤噪、降维、特征提取方法，实现海量光谱图像数据的有效挖掘，阐明抹茶粒度、色泽、功能成分等品质属性的显微光谱图像特征表达机制；研究数据预处理方法、特征融合方法和模型建立方法，最终构建抹茶品质的显微光谱图像信息融合评价模型，明确抹茶品质的显微高光谱成像快速无损检测机理，为构建抹茶品质的智能化无损检测新方法提供理论基础。

抹茶是一种超细微翠绿色茶粉，作为天然绿色食品，具有极高的附加值，但当前缺乏抹茶品质的客观、快速评价方法。项目提出抹茶品质的显微高光谱成像快速无损检测新思路，将传统高光谱成像检测技术从宏观领域拓展到微观领域。首先，研究了抹茶粒度、色泽、功能成分、感官品质的相关性，皮尔逊相关性分析表明，不同的理化指标与感官品质之间存在不同程度的相关性，部分理化指标之间也存在较强的相关性。这些为抹茶显微高光谱的光谱图像解析提供了基础。其次，研究了显微光谱图像的数据采集方式，优化了显微光谱图像信息的滤噪、降维和特征提取方法，实现了海量光谱图像数据的有效挖掘。同时研究了光谱图像数据的预处理方法、特征融合方法和模型建立方法，最终构建了抹茶粒度、功能成分、感官品质的光谱图像信息智能评价模型，预测结果理想。项目研究结果明确了抹茶品质的显微高光谱成像快速无损检测机理，为构建抹茶品质的智能化无损检测新方法提供了理论基础。