摘要
为了探讨近红外光谱分析技术直接检测烟丝常规化学成分的可行性,利用采集到的烟丝近红外光谱,采用偏最小二乘法(PLS)对测定烟丝样品的总糖、还原糖、烟碱、总氮、钾、氯含量数据与其相应的光谱数据进行拟合,建立了直接用于预测烟丝总糖、还原糖、烟碱、总氮、钾和氯含量的NIRS分析模型,平均相对偏差分别为2.2%,2.37%,1.82%,2.22%,3.68%和3.30%,变异系数均小于3.5%。该模型预测准确性和重现性较高,简便、快速,适用于批量烟丝检测,尤其是无需研磨烟粉,进一步减少了粉尘的污染,可以直接用于卷烟产品叶组配方质量监控。
NIRS technique is applied to the measurement of total sugar,reducing sugar,nicotine,total nitrogen,potassium and chlorine contents in tobacco cut.These contents can be measured with NIRS which is combined with PLS method.Analysis model has been established and the model parameters are optimized with internal cross test method.The model can directly predict the content of total sugar,reduced sugar,nicotine,total nitrogen,potassium and chlorine in tobacco cut.The average relative errors of total sugar,reduced sugar,nicotine,total nitrogen,potassium and chlorine are 2.2%,2.37%,1.82%,2.22%,3.68% and 3.30% respectively.Their variation coefficients are less than 3.5%.The repeatability and accuracy of the established model are tested.The results show that the model is accurate,reliable and convenient.The method is nondestructive to tobacco cut and decreases dust pollution.It can be used to directly detect the tobacco cut in batch and monitor the quality of cigarette blend.
出处
《激光与红外》
CAS
CSCD
北大核心
2012年第4期393-398,共6页
Laser & Infrared
基金
中国烟草总公司科技重点项目(No.110200801020)
云南中烟工业公司科技项目(No.2008YL07)
红云红河集团项目(No.HYHH2010YL02)资助
关键词
近红外
烟丝
偏最小二乘回归法
总糖
还原糖
总氮
烟碱
钾
氯
NIRS
tobacco cut
PLS
total sugar
reducing sugar
total nitrogen
nicotine
potassium
chlorine
