基于颜色特征和支持向量机的焦栀子炮制程度识别方法  

Identification method of Gardeniae Fructus Praeparatus processing degree based on color features and support vector machine

作　　者：章宇珍 王云[2] 程朋乐[1] 张村[2] ZHANG Yuzhen;WANG Yun;CHENG Pengle;ZHANG Cun(School of Technology,Beijig Foresery University,Beiing 100083,China;Institute of Chinese Materia Media,China Academy of Chinese Medical Sciences,Beijing 100700,China)

机构地区：[1]北京林业大学工学院,北京100083 [2]中国中医科学院中药研究所,北京100700

出　　处：《中华中医药杂志》2023年第8期3768-3772,共5页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家自然科学基金项目(No.81873010,No.82173979);中国中医科学院重大攻关项目科技创新工程(No.CI2021A04204)。

摘　　要：目的:量化焦栀子炮制过程中的颜色变化,探索基于颜色特征和支持向量机的焦栀子炮制程度识别方法。方法:构建不同炮制程度的栀子过程样品图像数据集,分割栀子过程样品图像的目标区域并提取其L、a、b颜色值,与栀子粉末的L、a、b进行相关性分析。将颜色特征进行主成分分析,并将特征向量输入支持向量机进行焦栀子炮制程度识别。结果:本方法的a和b值与栀子粉末的a和b值呈高度相关性,分别为0.99和-0.96,L值和Eab值相关性较弱,分别为0.60和0.77。主成分分析验证得到的分类模型良好,建立的支持向量机模型对焦栀子炮制程度的识别准确率达到100%。结论:基于颜色特征和支持向量机对焦栀子炮制程度的识别效果较优。Objective:To quantify the color change of Gardeniae Fructus Praeparatus during processing,and to explore a method for processing degree recognition of Gardeniae Fructus Praeparatus based on color features and support vector machine(SVM).Methods:The image data sets of Gardeniae Fructus process samples with different processing degrees were constructed,and the target region of Gardeniae Fructus process image was segmited and the L,a and b color values were extracted,and the correlation analysis was conducted between the L,a and b values of Gardeniae Fructus powder.The color features were analyzed by principal component analysis(PCA),and the feature vectors were input into SVM to identify the processing degree of Gardeniae Fructus Praeparatus.Results:The results showed that the a and b values of this method were highly correlated with the a and b values of Gardeniae Fructus powder(0.99 and-0.96,respectively),and the L and Eab values were weakly correlated(0.60 and 0.77,respectively).The classification model obtained by PCA verification was good,and the accuracy of the SVM model for identifying the processing degree of Gardeniae Fructus Praeparatus reached 100%.Conclusion:The processing degree of Gardeniae Fructus Praeparatus can be well recognized based on color features and SVM.

关 键 词：焦栀子 炮制 颜色特征 支持向量机 判别 机器学习 

分 类 号：R283[医药卫生—中药学]