Construction and application of tobacco quality similarity algorithm based on near-infrared spectroscopy

doi:10.14088/j.cnki.issn0439-8114.2026.05.027

Abstract

Abstract: To achieve the digital and quantitative evaluation of tobacco quality similarity, a similarity measurement algorithm based on near-infrared(NIR) spectroscopy was constructed. Based on the NIR spectral and sensory evaluation data of single-grade tobacco vertical formulation samples from 2022 and 2023, dimensionality reduction was conducted using principal component analysis (PCA). Subsequently, the cosine distance and Euclidean distance were effectively coupled by introducing a kernel function and the L2 norm to calculate and construct the similarity metric matrices and sensory score difference matrices among the samples, followed by a correlation analysis between the two. The results demonstrated that the similarity metric values of 10 and 9 samples in the two respective years exhibited a significant positive correlation (P<0.05) with their sensory score differences, accounting for 66.7% and 60.0% of the total samples in each year, respectively. This indicated that the metric value derived from the model maintained a good consistency with expert sensory evaluations; namely, a larger algorithmic metric value corresponded to a greater sensory difference and a lower similarity between samples. This study provided an effective novel technical approach and quantitative tool for tobacco raw material substitution, auxiliary design of tobacco blends, and stability evaluation of cigarette product quality.

Key words: near-infrared spectroscopy, tobacco quality, similarity metric, kernel function, L2norm, digital evaluation

CLC Number:

WANG Da-bin, ZHOU Xian-sheng, LIU Zhi-guang, PENG Fu-yu, YU Wei-song, QIU Jun. Construction and application of tobacco quality similarity algorithm based on near-infrared spectroscopy[J]. HUBEI AGRICULTURAL SCIENCES, 2026, 65(5): 172-178.

References

[1] ROWEIS S T, SAUL L K.Nonlinear dimensionality reduction by locally linear embedding[J]. Science,2000,290(5500): 2323-2326.
[2] HUANG Z H, XU X, ZUO L.Reinforcement learning with automatic basis construction based on isometric feature mapping[J]. Information sciences, 2014, 286: 209-227.
[3] GAN F, HOPKE P K, WANG J J.A spectral similarity measure using Bayesian statistics[J]. Analytica chimica acta, 2009, 635(2): 157-161.
[4] VAN DER MEER F. The effectiveness of spectral similarity measures for the analysis of hyperspectral imagery[J]. International journal of applied earth observation and geoinformation,2006, 8(1): 3-17.
[5] YU K, GUO G D, LI J, et al.Quantum algorithms for similarity measurement based on euclidean distance[J]. International journal of theoretical physics, 2020, 59(10): 3134-3144.
[6] HUO J, MA Y P, LU C T, et al.Mahalanobis distance based similarity regression learning of NIRS for quality assurance of tobacco product with different variable selection methods[J]. Spectrochimica acta part A: Molecular and biomolecular spectroscopy, 2021, 251: 119364.
[7] TU B, LI N Y, FANG L Y, et al.Hyperspectral image classification with a class-dependent spatial-spectral mixed metric[J]. Pattern recognition letters, 2019, 123: 16-22.
[8] 谢有超. 基于近红外光谱的卷烟相似性度量及数字化叶组配方设计研究[D].贵阳:贵州大学,2021.
[9] QIN Y H,DUAN K,WU L J,et al.Similarity measure method based on spectra subspace and locally linear embedding algorithm[J]. Infrared physics & technology, 2019, 100: 57-61.
[10] BI Y M, LI S T, ZHANG L L, et al.Quality evaluation of flue-cured tobacco by near infrared spectroscopy and spectral similarity method[J]. Spectrochimica acta part A: Molecular and biomolecular spectroscopy, 2019, 215: 398-404.
[11] 李石头,廖付,何文苗,等.基于近红外光谱相似的烟叶替代与卷烟配方维护[J].烟草科技,2020,53(2):88-93.
[12] TORCASSO M S, AI J T, CASELLA G, et al.Pseudo-spectral angle mapping for pixel and cell classification in highly multiplexed immunofluorescence images[J].Journal of medical imaging, 2024,11(6): 067502.
[13] ZHENG L Y,WEN Y,REN W,et al.Hyperspectral dark-field microscopy for pathogen detection based on spectral angle mapping[J]. Sensors and actuators B: chemical, 2022, 367: 132042.
[14] CHENG W W, SUN D W, PU H B, et al.Characterization of myofibrils cold structural deformation degrees of frozen pork using hyperspectral imaging coupled with spectral angle mapping algorithm[J]. Food chemistry, 2018, 239: 1001-1008.
[15] 杨榕彬,白洪涛,曹英晖,等.基于光谱信息散度-光谱角的自适应密度峰值聚类波段选择方法[J].吉林大学学报:信息科学版,2024,42(3):438-445.
[16] 姬旭升,李旭,万泽福,等.基于高空间分辨率卫星影像的新疆阿拉尔市棉花与枣树分类[J].中国农业科学,2019,52(6):997-1008.
[17] LIU D H, LIU G Y, LIU Z M.Some similarity measures of neutrosophic sets based on the euclidean distance and their application in medical diagnosis[J]. Computational and mathematical methods in medicine, 2018.7325938.
[18] 李宏达,李德成,曾荣.基于光谱相似性匹配的土壤有机碳估算[J].土壤学报,2021,58(5):1224-1233.
[19] 孙艳丽,张霞,帅通,等.光谱角—欧氏距离的高光谱图像辐射归一化[J].遥感学报,2015,19(4):618-626.
[20] 李蓓蓓,郭军伟,马骥,等.近红外光谱分析技术在烟草行业的应用研究进展[J].中国烟草学报,2025,31(3):148-160.
[21] 郑成旭,蒋佳磊,李庆祥,等.近红外光谱技术在烟草行业中的应用进展[J].中国无机分析化学,2026,16(2):206-218.
[22] ALTMAN N,KRZYWINSKI M.The curse(s) of dimensionality[J].Nature methods, 2018, 15:399-400.