摘要
应用傅里叶变换近红外(FT-NIR)光谱分析技术结合相似度匹配的定性分析方法,以来自云烟(软珍品)、云烟(紫)和云烟(红)3个牌号表香样品各20批的近红外透射光谱数据为基础,利用TQ统计软件分别建立以上牌号表香的相似度匹配鉴别模型,并对建模参数和模型的预测效果进行了评价。结果表明,云烟(软珍品)、云烟(紫)和云烟(红)3个牌号的表香模型对模型样品预测的相似度匹配值(SMV)均在95以上,而对非模型样品预测的SMV均非常低,说明若表香品质发生变化,就可以通过近红外光谱分析技术结合相似的匹配分析很好反映出来。因此,应用FT-NIR光谱分析技术可以对卷烟表香进行快速、无损和大批量检测,为卷烟表香的质量控制提供了新的思路和方法。
Similarity-match identifying models are established using the TQ statistical software based on the NIR transmission spectra of 20-batche-cigarette top flavor samples of 3 brands, Yunyan（ Soft Rarity） , Yunyan（Purple） and Yunyan（Red）. Established parameters and the identifying effects of the models are evaluated as well in this study. Results show that the Similarity Match Value（SMV） of modeled samples of the 3 brands is more than 95 ;while the SMV of the non-model samples is very low. It means that the application of FT-NIR Spectroscopy combining the qualitative analysis method of similarity match can recover the quality change of tobacco top flavor. Therefore, the application of FTNIR Spectroscopy can identify the tobacco top flavor rapidly, nondestructively and largely, and it provides new ideas and methods to control the quality of the tobacco top flavor.
出处
《激光与红外》
CAS
CSCD
北大核心
2013年第1期44-48,共5页
Laser & Infrared
基金
云南中烟工业公司项目(No.2011CP02)
国家自然科学基金项目(No.NSFC31160432)资助
关键词
近红外
质量控制
相似度匹配
卷烟表香
NIR spectroscopy
quality control
similarity match
tobacco top flavor
