基于跨感知仿生传感器融合的食品智能化感官检验研究

Human sensory test is an important tool in food quality test. A new idea for the study on intelligent test food quality by multi-information fusion from heteregenous biomimetic sensors, in this research proposal, is proposed on the basis of the principle of human sensory test. The biomimetic sensors will be designed and modified using some advanced function materials, and the sensors bionic mechanism in intelligent test quality will be explored. The biomimetic sensors of vision, olfactory and taste, which are similar to human eye, nose and tongue, will be used to obtain various sensors data. Raw data will be preprocessed through denoising and dimension reduction,and characteristic variables will be extracted from the preprocessed data.The characteristic variables will be correlated with the human sensory test results using artifical intelligent algorithm, and thus, the biomimetic sensors will be provided with the perception function as human sensory organs. Multi-information fusion from heteregenous biomimetic sensors will be achieved by mapping the sensors data into a higher dimension space according to an uniform rule and fabricating a joint vector. We will study the nonlinear mechanism between the sensors information and food sensory quality, and a nonlinear expert system (model), which is close to human perception behavior, will have to be built in order to achieve intelligent sensory testing of food quality. The achievements in this project will be expected to reveal the biomimetic mechanism of the corresponding sensors in food sensory test and clarify the law in multi-information fusion from the heteregenous sensors. Also, this study can provide the theoritical supports for the practical usage in food artifical intelligent test by means of instruments.

人工感官检验是食品质量检测的重要手段之一，本研究仿效人工感官检验机理，提出基于跨感知仿生传感器信息融合的食品智能化感官检验思路。筛选功能性材料，设计和修饰仿生传感器，探讨其仿生机理，并利用视、嗅、味等仿生传感器分别模拟人类眼、鼻、舌等器官获取各传感信息；研究利用数据净化、降维和特征信息提取方法，从海量数据中挖掘特征变量，设计智能学习算法将其与人工感官结果相耦合，赋予各仿生传感器相应的感知功能；将跨感知传感信息按相同法则映射到一个高维空间，并抽象出若干虚拟向量，实现跨感知传感信息融合；研究传感器特征信息与食品感官品质之间的非线性机理，通过非线性手段构建一个与人类感知行为相接近的专家系统（模型），以实现对食品品质的智能化感官检验。本项目研究可望在理论上揭示感知传感器在食品感官检验中的仿生机理，阐明跨感知传感器信息融合的规律，为实现食品仪器化、智能化感官检验最终能辅助人工感官检验奠定理论基础。

人工感官检验是食品质量检测的重要手段之一，本研究仿效人工感官检验机理，提出了基于跨感知仿生传感器信息融合的食品智能化感官检验新思路。筛选功能性材料，设计和修饰仿生传感器，探讨其仿生机理，并利用视、嗅、味等仿生传感器分别模拟人类眼、鼻、舌等器官获取各传感信息；利用数据净化、降维和特征信息提取方法，从海量数据中挖掘特征变量，设计智能学习算法并将其与人工感官结果相耦合，赋予各仿生传感器相应的感知功能；将跨感知传感信息按相同法则映射到一个高维空间，并抽象出若干虚拟向量，实现了跨感知传感信息融合；研究了传感器特征信息与食品感官品质之间的非线性机理，通过非线性手段构建一个与人类感知行为相接近的专家系统（模型），实现了食品品质的智能化感官检验。本项目研究在理论上揭示了感知传感器在食品感官检验中的仿生机理，阐明了跨感知传感器信息融合的规律，为实现食品仪器化、智能化感官检验提高了理论基础。