邓群芬　李源源　胡春琼　黄双

摘      要：为实现对假冒抗生素药物的快速识别，采集了7种抗生素106份样品的近红外漫反射光谱，采用一阶导数、标准归一化和Norris平滑对光谱预处理，选择9 300～4 000 cm^-1的光谱波段，通过相似度匹配和判别分析法建立抗生素的鉴别模型，并取样对该模型验证，预测集均能被正确的识别，取得了良好的效果。因此，近红外光谱结合判别分析法能快速、准确的对抗生素质量进行鉴别，为抗生素的真伪鉴别提供了一种新方法。

关  键  词：抗生素;近红外漫反射光谱;真伪鉴别

中图分类号：O 657       文献标识码： A       文章编号： 1671-0460（2020）06-1257-04

Rapid Identification of Antibiotics by Near-infrared Spectroscopy

DENG Qun-fen， LI Yuan-yuan， HU Chun-qiong， HUANG Shuang

（Key Lab of Process Analysis and Control of Sichuan Universities， Yibin University， Yibin Sichuan 644000， China）

Abstract： To achieve the rapid identification of counterfeit antibiotics， the NIR diffuse reflection spectra of 7 species 106 samples of antibiotics were collected in this study. Subsequently， the near-infrared spectra were pretreated with the first derivate， standard normal variate and Norris smoothing. After that， frequencies from 9 300～4 000 cm^-1 transmit wavelength range were selected. By means of the similarity matching and discriminant analysis， the identification model of antibiotics was established and verified by samples. The results showed that the prediction set could be correctly identified， and good results were achieved. In conclusion， the NIR spectroscopy combined with discriminant analysis can quickly and accurately identify these antibiotics， which provides a new approach for identifying antibiotics.

Key words： Antibiotics; Near-infrared diffuse reflection spectroscopy; Identification

抗生素在人们日常生活中扮演着非常重要角色。近年来，假药问题层出不穷，据权威调查，全球假药年销售额达到400亿美元，每年还在以约13%的速度增长。当下，传统假药，包含以非药品冒充药品、以低值药品冒充高价药品，越来越少了，取而代之的是未经批准私自在药品处方中增加或减少活性成分、用普通企业生产的药品冒充名牌药品等，本文主要研究的假药类型是真包装里装假药，通过检测药物本身对药物真伪进行鉴别，解决了扫描药品条形码无法鉴别药物真伪的问题。传统检测方法费时费力，当前用于假劣药品分析主要有化学法、色谱法、光谱法^[1]。光谱法特别是分子光谱法具有简单、快速、灵敏、无需样品处理、可原位检测、特征丰富等独特优势，具有天然的优势^[2]。

近红外（NIR）光谱是波長在780～2 526 nm区域的电磁辐射波^[3]，近红外光谱分析技术可用于无损和在线分析^[4]，被广泛应用于农业^{[ 5]}、食品^[6]、医药^[7]、石油化工^[8]等领域。……