在采用太赫兹时域光谱(THz-TDS)技术对9种常见毒品纯品和3种混合物进行实验研究,并得到它们在0. 2～2 THz频率范围的特征吸收光谱的基础上,用支持向量机(SVM)对毒品纯品和混合物的太赫兹吸收光谱进行了识别分类。用归一化预处理后的9种毒品和面粉的太赫兹吸收光谱训练libsvm模型。选用与训练光谱不同时间测得的毒品和混合物的太赫兹吸收光谱作为检测光谱,经过归一化预处理之后分别输入到训练好的libsvm模型中进行识别,识别率达100%。识别结果充分表明,用支持向量机可以实现对不同种类毒品的识别和鉴定,为太赫兹光谱技术用于毒品的检测和识别提供了另一种有效的方法。

On the base of absorption spectra in the range of 0.2～2 THz of nine illicit drugs and three mixed drugs obtained by using terahertz time-domain spectroscopy (THz-TDS) technique, the THz absorption spectra of different illicit drugs and mixed drugs were identified by support vector machines (SVM). Absorption spectra of the nine illicit drugs and flour, which were pretreated by normalized unit, were used to train libsvm program. Absorption spectra of the illicit drugs and the mixed drugs which were measured at different time and pretreated by normalized unit too, were identified by the libsvm and the identification rate was 100%. The results indicate that it is feasible to apply SVM to identification of illicit drugs, which provides an effective method in the secure inspection and identification for illicit drugs.