摘要
为了评价打叶复烤成品片烟的混配均匀性,提出了一种基于近红外光谱分析技术评价成品片烟混配均匀性的方法。在叶片复烤机出口收集配方模块成品片烟,采用近红外光谱仪测定样品的光谱,通过多元散射校正(MSC)结合S-G平滑滤波和一阶导数预处理方法将片烟光谱信息进行预处理,提高信噪比,然后利用主成分分析(PCA)提取光谱数据所包含的特征信息,得到成品片烟的马氏距离,并以马氏距离作为特征值,计算片烟混配均匀性系数,最后应用该方法评价不同批次配方模块片烟混配均匀性效果。结果表明:各批次成品片烟的变异系数为3.86%~8.33%,均匀度为91.67%~96.14%。这说明采用近红外光谱结合主成分分析和马氏距离的多元统计方法建立的打叶复烤片烟整体信息的混配均匀性评价方法是可行的。各批次成品片烟的均匀度都在91.00%以上,说明打叶复烤成品片烟质量的稳定性较好。因此,利用近红外光谱结合多元统计模型可以快速、准确地评价打叶复烤片烟掺配均匀性,为打叶复烤加工工序的精细化控制提供科学、全面的技术支持。
In order to evaluate the blending uniformity of the tobacco strips after threshing and redrying,a new method based on near-infrared spectroscopy analysis technology was proposed for quickly evaluating the blending uniformity of the tobacco strips.At the exit of the leaf redrying machine,the tobacco strips of formula modules was collected,the spectrum of the samples was measured by using a near-infrared spectrometer,and the spectral information of the tobacco strips was treated by multiplicative signal correction(MSC)combined with Savitzky-Golay(S-G)smooth filtering and first derivative pre-treatment method to improve the signal-to-noise ratio.Then principal component analysis(PCA)was used to extract the characteristic information contained in the spectral data and obtain the Mahalanobis distance of tobacco strips.Mahalanobis distance was used as eigenvalue to calculate the uniformity coefficient of the tobacco strips.Finally,this method was used to evaluate the blending uniformity of different batches of formula modules tobacco strips.The results showed that the coefficient of variation of each batch of tobacco strips was 3.86%-8.33%,the uniformity was 91.67%-96.14%.The results showed that the established method based on near-infrared spectroscopy combined with multivariate statistics method of principal component analysis and Mahalanobis distance was feasible to evaluate the blending uniformity of the overall information of tobacco strips after threshing and redrying.The uniformity of each batch of finished tobacco strips was above 91.00%,indicating that the quality stability of the tobacco strips after threshing and redrying was good.Therefore,near-infrared spectroscopy combined with multivariate statistical model could quickly and accurately evaluate the blending uniformity of tobacco strips after threshing and redrying,which could provide scientific and comprehensive technical supports for the fine control of threshing and redrying processing procedures.
作者
王玉真
邱昌桂
杨晶津
李思源
王发勇
华一崑
刘继辉
高占勇
汪显国
WANG Yu-zhen;QIU Chang-gui;YANG Jing-jin(HongyunHonghe Tobacco(Group)Co.,Ltd.,Kunming,Yunnan 650231;Yunnan Comtestor Co.,Ltd.,Kunming,Yunnan 650106)
出处
《安徽农业科学》
CAS
2022年第24期178-180,194,共4页
Journal of Anhui Agricultural Sciences
基金
红云红河烟草(集团)有限责任公司科技项目(HYHH2019-GY08)
红云红河烟草(集团)有限责任公司科技项目(HYHH2021GY02)。
关键词
近红外光谱
主成分分析
混合均匀性
打叶复烤
马氏距离
Near infrared spectroscopy
Principal component analysis
Blending uniformity
Threshing and redrying
Mahalanobis distance
