基于近红外光谱的烟叶油分识别研究

打开文本图片集

关键词近红外光谱;烟叶;油分档次;模型;分类效果中图分类号 TS4l+1 文献标识码A文章编号 0517-6611（2026）02-0173-06doi： 10.3969/j.issn.0517-6611.2026.02.031

AbstractOjectie]Toexploretheelatioshiptweenteearinfraredspectrumoftobacolavesandtheoilotentoftobacolea es，soastopovidfecefohddentifatioofbaoeafolottetdkingthgdeifatiooelfacc leafoilastheseachct，Otbacoafsaplesereedstetraingst，ndacnglasfcationoelsuing supportvectoracins，andomforests，ndulti-yerpereptroserecostuctedndfusedintoasingleclasifatioodel.76obc leafsaplesweedeeststtvaatetasiaofetivessoftsiooelResulasifoaccacfte fusionmodelonthetestsetreachesO.87.Onthedatasetwithaboringoilgrade，theprecisionandrecallofthemodelare.9OndO.93re spectively;ontedatastwitmeduolgade，eodelsprecisionatendcallateaOndO.O，spectivelyOnldata set，theprecisioatecallateofodelare8dO.8specivelyConcsioneasfcatioodelosds methodhaodsfatiftivesientigtedesofbaoleafoiasdoarfradspeftbaca

Key words Near-infrared spectroscopv：Tobacco leaves ： Oil grade：Model：Clasification effect

近年来近红外光谱分析技术（NIR）是一种绿色分析技术，凭借简单、高效、快速、低消耗、无污染等优势，在行业中受到广泛关注[]。（剩余6760字）