摘要
基于雷达图表示原理,对葛根粉的近红外光谱进行了鉴别研究。首先,利用近红外光谱技术获得了中药葛根粉在12000cm^(-1)~4000cm^(-1)范围内的近红外漫反射光谱作为指纹图谱,然后采用多层雷达图对指纹图谱逐步进行特征提取,最后用优选图特征进行分类研究。该方法在保证数据整体性的前提下,有很好的直观可视性,最小识别误差率达1.04%。研究结果表明该方法既是一种可行的葛根粉鉴别新方法,同时又可作为可视化中药指纹图谱模式分析新方法 。
According to the principle of radar graph representation, the near infrared spectra of Radix Pueraria power were identified. Firstly, by using the near infrared spectroscopy, the near infrared diffuse reflection spectra of Ra- dix Pueraria Power in the scope of 12000-1 -4000cm-1 were obtained to be as fingerprints. Secondly, the feature extraction for the fingerprints was performed gradually by using the multilayer radar graph. Finally, the classifica- tion was studied by using the optimized graph features. On the premise of guaranteeing the data integrity, this method has a good intuitive visibility, and its minimal recognition error rate can reach 1.04%. The experimental results showed that the study provided not only is a feasible new method for Radix Pueraria power identification, but also a new method for visualized pattern analysis of the fingerprints of traditional Chinese medicine
出处
《高技术通讯》
CAS
CSCD
北大核心
2015年第7期719-724,共6页
Chinese High Technology Letters
基金
中国博士后科学基金(2012M510722)
中国中医科学研究院(ZZ12001)
燕山大学博士基金(B692)
秦皇岛市科学技术研究与发展计划(201001A119)资助项目
关键词
葛根
近红外光谱
雷达图
中药
指纹图谱
特征提取
radix pueraria, near infrared spectra, radar graph, traditional Chinese medicine, fingerprint,feature extraction
