基于近红外光谱技术对鹿茸鉴别方法的研究

通过对鹿茸和伪劣鹿茸近红外光谱的分析,选取信息量丰富的7000~4500cm-1区段的光谱作为分析对象,用PCA结合Fisher识别法建立针对鹿茸真伪的识别函数,实现了对鹿茸真伪的准确识别,其中定标集样本的识别准确率达到了98.3%,外部鹿茸样本验证集识别准确率达到了93.5%。

Study on Using Near-Infrared Spectroscopy Technique to Identify Velvet

The velvet antler and fake antler was analysed by near - infrared specturm, thus that the rich spectroscpic data of 7000 - 4500 cm - 1 section was selected. Otherwise, we also developed the identification function for the authenticity of the velvet antler by an integrating method including PCA and Fisher recognition method. It allows accurate identification of the authenticity, the accuracy rate reached 98.3% in the recognition of the calibration set, and that reached 93.5% in the recognition of the validation set samples outside pilose antler..