手持近红外光谱法快速定量检测初烤烟叶中的尼古丁含量

Rapid Quantitative Detection of Nicotine Content in Freshly Roasted Tobacco Leaves Using Handheld Near-Infrared Spectroscopy

【目的】近红外光谱技术具有使用简便、无损、快速、准确的优点,使用手持近红外方法在现场对原烟片进行无损检测对烟片的现场管控分级等具有重要意义。【方法】本研究采用微机电(MEMS)手持近红外光谱技术,基于不同产地等级、类型的398个初烤烟叶样品进行建模,烟叶不做任何处理直接扫描,对所得光谱进行一阶导数、平滑以及范围归一化处理等,并采用偏最小二乘法(PLS)建立了初烤烟片的尼古丁预测模型进行效果评估。【结果】该模型线性良好,R2达到0.78,RMSEC值为0.44;且模型预测效果不错,绝对偏差在0.1%以内,相对偏差在5%以内,精度完全满足现场无处理快速检测烟片的要求,可以实际使用。【结论】本研究能够为烟片的现场无处理快速检测、烟片质控以及烟片的现场分级,提供切实、可行的指导和借鉴,后续可对水分、总糖、总氮等其他相关指标进行建模,完善手持近红外光谱技术对于现场烟片检测的使用效果。

[Objective]Near infrared spectroscopy technology has the advantages of simple,non-destructive,fast and accurate use.The use of handheld near-infrared methods for non-destructive testing of original tobacco slices on site is of great significance for the on-site control and grading of tobacco slices.[Method]This study used MEMS handheld near-infrared spectroscopy technology to model 398 samples of freshly cured tobacco leaves from different production areas,grades and types.The tobacco leaves were directly scanned without any processing,and the obtained spectra were subjected to first-order derivative,smoothing,and range normalization processing.Partial least squares(PLS)method was used to establish a nicotine prediction model for freshly cured tobacco leaves for effectiveness evaluation.[Result]The model has good linearity,with an R2 of 0.78 and an RMSEC value of 0.44.And the model prediction effect is good,with an absolute deviation of less than 0.1%and a relative deviation of less than 5%.The accuracy fully meets the requirements of on-site rapid detection of tobacco without processing,and can be used in practice.[Conclusion]This study can provide practical and feasible guidance and reference for on-site rapid detection,quality control,and grading of tobacco without processing.In the future,other related indicators such as moisture,total sugar,and total nitrogen can be modeled to improve the effectiveness of handheld near-infrared spectroscopy for on-site tobacco detection.