基于近红外光谱技术的毒品现场快速溯源研究  

On-site Rapid Source Identification of Drugs Based on Near-infrared Spectroscopy Technology

作　　者：张建强 李帆 杨啟富 胡竣勋 倪春明 张馨予[2] ZHANG Jian-qiang;LI Fan;YANG Qi-fu;HU Jun-xun;NI Chun-ming;ZHANG Xin-yu(Institute of Criminal Investigation,Yunnan Police College,Kunming 650021,China;Faculty of Science,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]云南警官学院刑事侦查学院,云南昆明650021 [2]昆明理工大学理学院,云南昆明650500

出　　处：《分析测试学报》2022年第7期1007-1013,共7页Journal of Instrumental Analysis

基　　金：云南警官学院物证光谱技术省创新团队(202105AE160007);公安部科技强警基础研究项目(2020GABJC41,2019GABJC40);云南省高校物证光谱技术重点实验室;云南省刑事科学技术重点实验室开放课题(2020zz02);云南警官学院校级科研项目(21A028);大学生创新创业项目(202011392004)。

摘　　要：毒品的日益泛滥和快速更新对缉毒部门进行现场识别和快速检测带来了巨大挑战,寻求快速、准确、低成本的现场检测方法对毒品的有效管控和案件侦破具有重要意义。该文通过使用手持式近红外光谱仪,结合粒子群优化-极限学习机(PSO-ELM)算法建立了冰毒和海洛因的现场案件快速溯源模型,并与传统的线性判别分析(LDA)算法和支持向量机(SVM)算法进行比较。结果表明:相对于传统的分类模型,该文所建立的PSO-ELM案件溯源模型能获得更好的分类结果和更高的计算效率。该方法能够实现毒品的现场快速溯源,为禁毒实战中的案件侦破提供制毒、贩毒线索。The increasing proliferation and quick upstate of the synthetic new drugs have brought huge challenges to the anti-narcotics units in field identification and rapid detection.It is of positive significance to seek a fast,accurate and low-cost on-site detection method for the effective control of the new drugs.In the paper,a novel particle swarm optimization-extreme learning machine(PSO-ELM)algorithm combined with a hand-held near-infrared(NIR)spectroscopy was applied to build the source recognition models for methamphetamine and heroin drugs,which was also compared with the traditional linear discriminant analysis(LDA)algorithm and support vector machine(SVM)algorithm.The experimental results showed that the PSO-ELM source recognition models could obtain the best classification results and computational efficiency.The proposed method could recognize the different sources of the drugs,provide some valuable clues for determining the nature of the case in the prohibition of actual combat.

关 键 词：近红外光谱(NIR) 毒品 极限学习机(ELM) 溯源 

分 类 号：O657.33[理学—分析化学]