基于视觉和嗅觉传感器数据融合技术的红茶通氧发酵过程在线监测研究

Herein, we proposed a new idea that the real-time monitoring of black tea oxidization fermentation could be achieved using data fusion based on vision and olfaction sensors, because black tea fermentation is a complex and dynamic process.A intelligent solid state fermentation tank will be designed for the experiment of black tea oxidization fermentation. The mechanism of color changing and aroma changing will be focused on during black tea oxidization fermentation. A portable visible diffuse reflectance spectroscopy system, as the vision sensor in this work, wiil be developed and used for data acquisition related to the changing of tea color during black tea oxidization fermentation. A olfaction system based on colorimetric sensors will be developed and used for data acquisition related to the changing of tea aroma during black tea oxidization fermentation. Raw data will be preprocessed using denoising and dimension reduction algorithms,and characteristic variables will be extracted from the preprocessed data. Data fusion from heteregenous sensors will be achieved by mapping the multi-dimensional sensors data into a higher dimension space by a rule to form a joint vector.Finally, we constucted a model to achieve the real-time monitoring of black tea fermentation process by selecting a suitable pattern recognition. By this project, we shall be expected to construct the theoretical framework for the information-based and intelligent black tea fermentation technology and equipment.

针对红茶发酵过程的复杂性和时变性，研究提出基于多传感器信息融合技术的红茶通氧发酵过程在线监测思路。设计一套智能化固态发酵罐用于红茶通氧发酵实验平台，利用专家经验监测和理化分析手段，研究红茶通氧发酵过程中叶面色泽和香气的变化及其与发酵适度的相关性，为在线监测研究提供参考标准；设计一套便携式可见光漫反射光谱系统作为视觉传感器，实时获取红茶发酵过程中叶面色泽变化信息；筛选色敏材料，设计一套色敏型嗅觉传感器系统，实时获取红茶发酵过程中香气变化信息；精选数据挖掘算法对原始信号进行校正、滤噪和降维，从海量数据中提取特征向量，并将其按统一法则映射到一个高维特征空间，通过构建虚拟向量，完成视觉和嗅觉征向量的数据融合；对照专家经验监测和理化分析结果形成的知识库，优选合适的模式识别方法构建判别模型，实现对红茶发酵过程中发酵适度的在线监测。本项目研究将为最终实现红茶发酵工艺的信息化和装备的智能化奠定理论基础。

针对红茶发酵过程的复杂性和时变性，提出了基于多传感器信息融合技术的红茶通氧发酵过程在线监测思路。设计了一套智能化固态发酵罐用于红茶通氧发酵实验平台，利用专家经验监测和理化分析手段，研究了红茶通氧发酵过程中叶面色泽和香气的变化及其与发酵适度的相关性，为在线监测研究提供了参考标准；设计了一套便携式可见光漫反射光谱系统作为视觉传感器，实时获取红茶发酵过程中叶面色泽变化信息；针对性筛选了色敏材料，明确了色敏材料的显色机理，并进行了针对性重构，开发了一种灵敏度高、特异性强的传感芯片，在此基础上，设计了一套色敏型嗅觉传感器监测系统，实时获取了红茶发酵过程中香气变化信息；精选数据挖掘算法对原始信号进行校正、滤噪和降维，从海量数据中提取特征向量，并将其按统一法则映射到一个高维特征空间，通过构建虚拟向量，完成视觉和嗅觉征向量的数据融合；在此基础上，对照专家经验监测和理化分析结果形成的知识库，优选合适的模式识别方法构建判别模型，实现了红茶发酵过程中发酵适度的在线监测。本项目研究将为最终实现红茶发酵工艺的信息化和装备的智能化奠定理论基础。