基于太赫兹光谱检测的危险品模糊识别研究

Fuzzy Recognition Research on Explosive and Illegal Drug Based on Terahertz Spectroscopy Detection

在采用太赫兹时域光谱（THz-TDS）技术对7种常见危险品进行实验研究并得到它们在0.2～2.6 THz频率范围的特征吸收谱的基础上，用模糊识别方法对7种常见危险品的太赫兹特征吸收谱进行了训练和识别。将爆炸物RDX（黑索金）、γ-HNIW、DNT（2,4-二硝基甲苯）及毒品MA（甲基苯丙胺）、氯胺酮等的不同特征吸收峰作为模糊聚类分析数据源，利用相关系数法建立模糊相似矩阵，借助传递闭包法获得模糊等价矩阵，形成标准太赫兹吸收光谱模型库，采用极差分析进行数据预处理，计算海明贴近度，识别待检物品为隐蔽在军服后的RDX和毒品MDA（替苯丙胺）。研究结果表明，由分子间相互作用或声子共振模式引起的不同特征吸收是确定危险品类型和种类的依据，用模糊识别可以实现对不同种类危险物品的识别和鉴定，为太赫兹光谱技术用于危险品的检测和识别提供了一种新的有效方法。

On the base of absorption spectra in the range from 0.2 to 2.6 THz of seven common dangerous goods using terahertz time-domain spectroscopy （THz-TDS） technique, the THz absorption spectra of seven dangerous goods were identified success- fully by fuzzy recognition. Using different feature absorption peaks of explosive and illegal drugs such as RDX, γ-HNIW, DNT, MA and Ketamine as data source of fuzzy cluster analysis, fuzzy similar matrix was built by correlation coefficient, fuzzy equiva- lent matrix was obtained by the method of transitive closure, standard model bank of THz absorption spectra was formed, and data were preprocessed by range analysis and compute Hamming approach degree. It was concluded that uninspected goods were RI）X hidden behind the uniform and MDA. The research indicated that different feature absorption by interactions between mole- cules and phonon resonance mode is the basis for determining the type and category of dangerous goods, it is feasible to apply fuzzy recognition to the identification of dangerous goods, providing an effective new method for the secure inspection and identi- fication of threats using THz TDS technique.