目的：探索基于支持向量机（SVM）的“内外结合”中药质量控制新模式。方法：利用仿生嗅觉、视觉及味觉系统采集中药

材及其饮片的性状、颜色、气味、味道等外部信息，再利用SVM与中药材中化学成分的含量、水分、灰分、浸出物等内部信息建立关

联数据库，建立基于SVM的中药整体质量控制新模式。选用来自四川、重庆、湖北和云南的黄连作为分析对象，分别采集其内外

部和产地信息后对该系统进行有效性验证。结果与结论：内部特征、外部特征、“内外结合”全部特征这3 种识别方法的识别率均

大于90％，且“内外结合”全部特征的识别率较单一内部或外部特征高。选用SVM法对得到的内外数据建立关联和特征模型，可

使传统“辨状论质”经验得以现代“察外知内”的再现，从而在传统经验的基础上建立更加规范、准确、客观的中药质量整体控制新

模式。

OBJECTIVE：To explore“internal and external combined”new traditional Chinese medicine（TCM）quality control

mode based on support vector machines（SVM）. METHODS：External data were collected with bionic olfaction，vision and taste

system，such as property，color，smell，taste；internal data as the content of effective ingredients，moisture，ash content and extracts

were collected to establish relational database with SVM；new TCM quality control mode was established on the basis of

SVM. Coptis chinensis from Sichuan，Chongqing，Hubei and Yunnan were selected as subject. External and internal data and producing

area information were collected to validate the effectiveness of the system. RESULTS & CONCLUSIONS：The recognition

rate of 3 methods are all higher than 90％，including internal characteristics，external characteristics and whole internal and external

characteristics；the recognition rate of whole internal and external characteristics is higher than that of internal characteristics or

external characteristics. Relational and characteristics model of internal and external data are established with SVM method，and realize

traditional experience of“identifying quality from appearance”through modern method to“knowing internal quality from external

characteristics”；so，more standardized，accurate and objective new TCM quality control mode will be established on the basis

of traditional experience.