基于智能感官与多源信息融合技术的香附炮制程度快速辨识方法研究

目的 基于多源信息融合技术，整合传统的中药性状鉴别，建立香附Cyperi Rhizoma炮制程度快速辨识方法，为香附质量评价标准的制定和炮制过程质量控制的应用研究提供新思路、新方法。方法 选取6个产地的生香附饮片，采用醋炙法炮制，每隔3 min取样，得到72批香附炮制过程样品。然后基于色差仪、电子鼻和近红外光谱（near infrared spectrum,NIRS）技术获取上述样品的智能感官信息和NIRS数据，利用主成分分析-判别分析（principal component analysis-discriminant analysis,PCA-DA）、偏最小二乘-判别分析（partial least squares-discriminant analysis,PLS-DA）、正交偏最小二乘-判别分析（orthogonal partial least squares-discriminant analysis,OPLS-DA）方法、Lasso回归分析、遗传算法（genetic algorithm,GA）-反向传播（back propagation,BP）、神经网络算法（GA-BP ne...

Research on rapid identification method of processing degree of Cyperi Rhizoma based on intelligent sense and multi-source information fusion technology

Objective Based on multi-source information fusion technology, integrating traditional Chinese medicine character identification, a rapid identification method of processing degree of Xiangfu (Cyperi Rhizoma) was established, which provided a new idea and a new method for the formulation of quality evaluation criteria and the application research of processing quality control. Methods A total of 72 batches of fragrant decoction pieces from six regions were processed with vinegar and sampled at a interval of 3 min. Then, the intelligent sensory information and near-infrared spectral data of the above samples were obtained based on the color difference meter, electronic nose and near-infrared spectrum (NIRS), and the principal component analysis-discriminant analysis (PCA-DA), partial least squares-discriminant analysis (PLS-DA), orthogonal partial least squares-discriminant analysis (OPLS-DA), Lasso regression analysis, genetic algorithm (GA)-back propagation (BP) neural network algorithm (GA-BPNNA) and other stoichiometric methods were used, then the processing degree identification method was established based on single source information and multi-source information fusion, respectively. Results The processing degree identification model based on single source of color difference meter, electronic nose and NISR could not accurately identify the four types of processed products of Cyperi Rhizoma, while the processing degree identification model based on two kinds of intelligent senses and multi-source information fusion technology could quickly and accurately identify the four types of processed products of Cyperi Rhizoma, with an accuracy of more than 0.93, and the model classification and prediction effect were good. Conclusion The processing degree identification model of Cyperi Rhizoma based on two kinds of intelligent senses and multi-source information fusion technology can identify the processing degree of Cyperi Rhizoma more accurately, further improve the prediction accuracy, and provide a reference for the rapid identification of the processing degree of Cyperi Rhizoma and other traditional Chinese medicine.