基于云端-互联便携式近红外技术现场快检西红花真伪

doi:10.3964/j.issn.1000-0593(2020)10-3029-09

摘要
参考文献
相关文章 (15)

全文: PDF (5753 KB)
输出: BibTeX | EndNote (RIS)

摘要：利用云端-互联便携式近红外技术结合化学计量学对名贵药材西红花与其常见伪品（红花、玉米须、莲须、菊花、纸浆）和掺伪品进行现场快速真伪鉴别及掺伪量的定量预测。用移动手机控制的PV500R-I便携式近红外仪采集西红花与其伪品和掺伪品光谱数据。对原始光谱数据进行一阶导，二阶导，三阶导，标准正态变量转换和光散射校正前处理。采用偏最小二乘判别分析分步建立西红花与其伪品、西红花与其掺伪品鉴别模型。结果表明，一个最优模型可将西红花与其五类伪品彼此完全区分；两个最优模型分步区分西红花与其五类掺伪品，外部预测准确率最低为93%，西红花掺入红花、玉米须、莲须、菊花和纸浆的识别水平分别为0.5%，0.5%，4.0%，0.5%和0.5%。采用偏最小二乘回归对五类西红花掺伪品的掺伪量建立定量预测模型，五个最优模型的外部预测相关系数范围为0.920～0.999，RMSEP范围为0.005～0.044，当西红花掺入红花、菊花、莲须、纸浆和玉米须的掺伪量大于8%时，其外部预测相对误差分别低于8%，8%，3%，10%和5%，表明最终模型能较好地预测西红花掺伪品的掺伪量。基于云端-互联便携式近红外光谱技术所建立的西红花真伪鉴别方法和掺伪品掺伪量预测方法快速准确，经济环保，能满足西红花现场快速无损伤真伪鉴别要求。

关键词：云端-互联便携式近红外技术；化学计量学；西红花；伪品；掺伪品

Abstract：The use of cloud-connected infrared spectroscopy technology combined with chemometrics to identify the rarest saffron and its commonly encountered adulterants (carthami flos, corn silk, nelumbinis stamen, chrysanthemi flos, pulp) and adulterated saffron, and quantitative determination of adulterant in saffron. Near-infrared spectra of saffron, adulterants, and adulterated saffron were collected by using the PV500R-I portable near-infrared instrument controlled by mobile phone. The first derivative, second derivative, third derivative, standard normal variable transformation and multiplicative scatter correction are used to preprocess the original spectral data. Partial Least Squares Discrimination Analysis was used to establish the identification model of saffron and its adulterant, and saffron and adulterated saffron. The results show that an optimal recognition model can distinguish saffron and its five kinds of adulterant from each other completely; the lowest 93% external prediction accuracy of saffron and five kinds of samples of adulterated saffron can be achieved step-by-step by two optimal recognition models, and the adulteration recognition level of saffron mixed with carthami flos, corn silk, nelumbinis stamen, chrysanthemi flos and pulp are 0.5%, 0.5%, 4.0%, 0.5% and 0.5%, respectively. Partial least squares regression was used to establish quantitative prediction models for the five kinds adulterant in saffron. The external prediction correlation coefficient range of the final model was 0.920～0.999, and RMSEP range was 0.005～0.044, and when the saffron mixed with carthami flos, chrysanthemi flos, nelumbinis stamen, pulp and corn silk are more than 8%, its external prediction relative error is less than 8%, 8%, 3%, 10% and 5% respectively, which indicated that the quantitative prediction model could be used to predict the amount of adulterant in saffron. To sum up, the identification method based on cloud connected portable near-infrared spectroscopy and the prediction method of the amount of adulterant is fast and accurate, economic and environmental protection, and can meet the requirements of quick and non-destructive identification of saffron on site.

Key words：Cloud-connected portable near-infrared technology; Chemometrics; Saffron; Adulterant; Adulterated samples

收稿日期: 2020-05-11 修订日期: 2020-07-15

中图分类号:

O657.3

基金资助: 四川省重大科技专项（2018TZDZX0007）资助

通讯作者: 严铸云 E-mail: yanzhuyun@cdutcm.edu.cn

作者简介: 李庆，1980年生，成都中医药大学、食品药品检验研究院在读博士研究生 e-mail: liqqin2001@outlook.com

引用本文:

李庆，闫晓剑，赵魁，李蘭，彭善贵，罗霄，文永盛，严铸云. 基于云端-互联便携式近红外技术现场快检西红花真伪[J]. 光谱学与光谱分析, 2020, 40(10): 3029-3037.
LI Qing, YAN Xiao-jian, ZHAO Kui, LI Lan, PENG Shan-gui, LUO Xiao, WEN Yong-sheng, YAN Zhu-yun. Fast Inspection of Saffron on the Spot Based on Cloud-Connected Portable Near-Infrared Technology. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2020, 40(10): 3029-3037.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2020)10-3029-09 或 https://www.gpxygpfx.com/CN/Y2020/V40/I10/3029