| 485 | 19 | 20 |
| 下载次数 | 被引频次 | 阅读次数 |
【目的】为提高烤烟的分类正确率。【方法】利用高光谱成像系统采集烤烟样本，采用Savitzky-Golay卷积平滑法（SG）、多元散射校正（MSC）、一阶导数法（FD）的组合方法对数据进行预处理。通过连续投影算法（SPA）选择特征波长，利用灰度共生矩阵（GLCM）选择烤烟的纹理特征，将纹理特征与光谱特征归一化处理后进行融合，利用邻近算法(KNN)、随机森林(RF)、支持向量机(SVM)、朴素贝叶斯（NB）验证烤烟分类效果。【结果】预处理后的全波长数据分类正确率较预处理前有所提升；利用SPA选择特征波段进行分类，正确率下降；高光谱融合纹理特征后进行分类，分类效果显著提升。【结论】基于高光谱与纹理融合可准确、有效地对烤烟进行无损分类，为量化烤烟分类提供了可行方法。
Abstract：[Objective]The purpose of this paper is to improve the classification accuracy of flue-cured tobacco. [Methods]A hyperspectral imaging system was used to collect flue-cured tobacco sample. The combination method of Savitzky-Golay(SG) smoothing filter,multiplicative scatter correction(MSC) and the first derivative(FD) was used to preprocess the data of flue-cured tobacco. Feature wavelength was selected through successive projections algorithm(SPA). The texture features of flue-cured tobacco selected by gray-level co-occurrence matrix(GLCM), and then the texture features and spectral features were normalized and fused. The classification effect of flue-cured tobacco was verified by using k-nearest neighbor(KNN), random forests(RF), support vector machine(SVM) and naive bayes(NB). [Results]The results showed that the classification accuracy of full-band hyperspectral data after preprocessing was improved compared with that before processing. 25 features bands were selected by SPA, and the classification accuracy using features bands was decreased. After the fusion of hyperspectral and texture features, the classification effect was improved significantly. [Conclusion] The proposed method can accurately and effectively classify flue-cured tobacco without damage through fusion of hyperspectral and texture,which provides a feasible method for quantifying the classification of flue-cured tobacco.
[1]方世诚.基于光谱与色度的烤烟分类方法研究[D].贵阳：贵州大学，2016.FANG Shicheng. Study on classification method of flue cured tobacco based on spectrum and chroma[D]. Guiyang:Guizhou University,2016.
[2]中华人民共和国国家质量监督检验检疫总局.烤烟：GB2635-1992[S].北京：中国标准出版社，1992.General administration of quality supervision, inspection and quarantine of the People’s Republic of China. Flue-Cured Tobacco:GB 2635-1992[S]. Beijing:China Standards Press,1992.
[3]陆江明，范婷婷，穆青爽，等.融合光谱与纹理特征的龙井茶等级无损识别[J].现代食品科技，2021, 37(3):301-307, 249.LU Jiangming, FAN Tingting, MU Qingshuang, et al.Nondestructive identification of longjing Tea Grade by fusing spectral and textural feature[J]. Modern Food Science and Technology, 2021, 37(3):301-307, 249.
[4]郭婷.不同钾水平和成熟度烟草高光谱特征及其品质估测模型研究[D].长沙：湖南农业大学，2019.GUO Ting, Hyperspectral characteristics and estimating model of leaf quality content in tobacco based on potassium and maturity treatments[D]. Changsha:Hunan University,2019.
[5]杨艳东，贾方方，刘新源，等.烤烟叶片氯密度高光谱预测模型的建立[J].河南农业科学，2019,48(5):155-160.YANG Yandong, JIA Fangfang, LIU Xinyuan, et al. Construction of hyperspectral prediction model for chlorine density of flue-cured tobacco leaves[J]. Journal of Henan Agricultural Sciences, 2019,48(5):155-160.
[6] Mc Clure W F. Spectral characteristucs of tobacco in the near infrared region from 0.6-2.6 microns[J]. Tobacco Sci, 1968(12):232-235.
[7]宾俊，周冀衡，范伟，等.基于NIR技术和ELM的烤烟烟叶自动分类[J].中国烟草学报，2017,23(2):60-68.BIN Jun, ZHOU Jiheng, FAN Wei, et al. Automatic grading of flue-cured tobacco leaves based on NIR technology and extreme learning machine algorithm[J]. Acta Tabacaria Sinica, 2017, 23(2):60-68.
[8]赵凡，闫昭如，薛建新，等.高光谱无损识别野生和种植黑枸杞[J].光谱学与光谱分析，2021,41(1):201-205.ZHAO Fan, YAN Zhaoru, XUE Jianxin, et al. Identification of wild black and cultivated goji berries by hyperspectral image[J].Spectroscopy and Spectral Analysis, 2021, 41(1):201-205.
[9]刘立新，李梦珠，赵志刚，等.高光谱成像技术在生物医学中的应用进展[J].中国激光，2018,45(2):214-223.LIU Lixin, LI Mengzhu, ZHAO Zhigang, et al. Recent advances of hyperspectral imaging application in biomedicine[J]. Chinese Journal of Lasers, 2018,45(2):214-223.
[10]张龙，马啸宇，王锐亮，等.高光谱成像技术在烟叶和杂物分类中的应用[J].烟草科技，2020,53(8):72-78.ZHANG Long, MA Xiaoyu, WANG Ruiliang, et al. Application of hyperspectral imaging technology to tobacco leaf and foreign matter sorting[J]. Tobacco Science&Technology, 2020, 53(8):72-78.
[11]贾方方，马新明，李春明，等.不同水分处理对烟草叶片高光谱及红边特征的影响[J].中国生态农业学报, 2011, 19(6):1330-1335.JIA Fangfang, MA Xinming, LI Chunming, et al. Effect of water condition on hyperspectral and red-edge characteristics of tobacco leaf[J]. Chinese Journal of Eco-Agriculture,2011,19(6):1330-1335.
[12]陈朋.烟叶纹理表面的视觉检测技术研究与应用[D].贵阳：贵州大学，2017.CHEN Peng. Research and application of visual inspection technology for texture of tabacco surface[D]. Guiyang:Guizhou University, 2017.
[13]郑守国，翁士状，刘瑜凡，等.融合高光谱成像多类特征的名优牛肉种类鉴别[J].激光杂志，2021,42(8):57-61.ZHENG Shouguo, WENG Shizhuang, LIU Yufan, et al. Identification of famous beef species using multiple futures of hyperspectral imaging[J]. Laser Journal, 2021,42(8):57-61.
[14]李海杰.基于机器视觉的烟草异物检测和烟叶分类分级方法研究[D].南京：南京航空航天大学，2016.LI Haijie. Research on methods of tobacco foreign material detection and tobacco leaves grading based on machine vision[D].Nanjing:Nanjing University of Aeronautics and Astronautics,2016.
[15]赵鹏，韩金城，王承琨.基于I-BGLAM纹理和光谱融合的高光谱显微成像木材树种分类[J].光谱学与光谱分析, 2021, 41(2):599-605.ZHAO Peng, HAN Jincheng, WANG Chengkun. Wood species classification with microscopic hyper-spectral imaging based on I-BGLAM texture and spectral fusion[J]. Spectroscopy and Spectral Analysis,2021,41(2):599-605.
[16]黄蒂云，李景彬，尤佳，等.基于高光谱技术融合图像信息的脱绒棉种品种分类检测研究[J].光谱学与光谱分析，2018, 38(7):2227-2232.HUANG Diyun, LI Jingbin, YOU Jia, et al. The classification of delinted cottonseeds varieties by fusing image information based on hyperspectral image technology[J]. Spectroscpy and Spectral Analysis, 2018, 38(7):2227-2232.
[17]何嘉琳，乔春燕，李冬冬，等.可见近红外高光谱成像技术对灵武长枣VC含量的无损检测方法[J].食品科学，2018,39(6):194-199.HE Jialin, QIAO Chunyan, LI Dongdong, et al. Non-destructive detection of vitamin c content in “lingwu changzao” jujubes using visible near infrared hyperspectral imaging[J]. Food Science, 2018,39(6):194-199.
[18]张晟翀，李宇海.红外高光谱数据鉴别技术研究[J].光电技术应用，2019, 34(2):27-33.ZHANG Shengchong, LI Yuhai. Research on infrared hyperspectral data identification technology[J]. Electro-Optic Technology Application, 2019, 34(2):27-33.
[19]孟庆龙，张艳，尚静.光纤光谱结合模式识别无损检测苹果表面疤痕[J].激光技术，2019, 43(5):86-90.MENG Qinglong, ZHANG Yan, SHANG Jing. Nondestructive detection of apple defect combining optical fiber spectra with pattern recognition[J]. Laser Technology, 2019, 43(5):86-90.
[20]芦兵，孙俊，杨宁，等.基于荧光透射谱和高光谱图像纹理的茶叶病害预测研究[J].光谱学与光谱分析, 2019, 39(8):2515-2521.LU Bing, SUN Jun, YANG Ning, et al. Prediction of tea diseases based on fluorescence transmission spectrum and texture of hyperspectral image[J]. Spectroscopy and Spectral Analysis, 2019,39(8):2515-2521.
[21]孙婷，田建平，胡新军，等.基于高光谱成像技术的酿酒高粱品种分类[J].食品与发酵工业，2021,47(05):186-192.SUN Ting, TIAN Jianping, HU Xinjun, et al. Classification of liquor sorghum varieties based on hyperspectral imaging technology[J].Food and Fermentation Industries, 2021, 47(5):186-192.
[22]高蕊.随机森林算法改进及其在医疗诊断系统中的应用[D].南京：南京邮电大学，2020.GAO Rui. Improvement of random forest algorithm and its application in medical diagnosis system[D]. Nanjing:Nanjing University of Posts and Telecommunications, 2020.
[23]王靖会，程娇娇，刘洋，等.基于高光谱成像技术鉴别大米品种[J].中国农业科技导报，2021, 23(9):121-128.WANG Jinghui, CHENG Jiaojiao, LIU Yang, et al. Identification of rice variety based on hyperspectral imaging technology[J].Journal of Agricultural Science and Technology, 2121, 23(9):121-128.
[24]王智宏，刘杰，王婧茹，等.数据预处理方法对油页岩含油率近红外光谱分析的影响[J].吉林大学学报(工学版)，2013, 43(04):1017-1022.WANG Zhihong, LIU Jie, WANG Jingru, et al. Influence of data preprocessing on NIRS analysis for oil yield of oil shale[J]. Journal of Jilin Univesity(Engineering and Technology Edition), 2013,43(04):1017-1022.
[25]王阳.基于朴素贝叶斯的高光谱矿物识别[D].长春：吉林大学，2014.WANG Yang, Mineral identification by hyperspectral based on data mining[D]. Changchun:Jilin University, 2014.
[26]刘爽.小麦赤霉病高光谱信息多循环提取及组合式识别研究[D].长春：中国科学院大学(中国科学院长春光学精密机械与物理研究所)，2019.LIU Shuang. Research on multi-cycle extraction and combination identification of hyperspectral information of fusarium head blight in wheat[D]. Changchun:University of Chinese Academy of Sciences(Changchun Institute of Optics,Fine Mechanics and Physics), 2019.
基本信息:
DOI：10.16472/j.chinatobacco.2021.T0139
中图分类号:TP751;TS44
引用信息:
[1]张慧,张文伟,张永毅等.基于高光谱与纹理融合的烤烟分类方法研究[J].中国烟草学报,2022,28(03):72-80.DOI:10.16472/j.chinatobacco.2021.T0139.
基金信息:
中国烟草总公司福建省公司科技项目“烟草收购全程质量追溯管理模式构建与应用”（闽烟司[2022]5号）;中国烟草总公司福建省公司科技计划项目“鲜烟成熟度智能化检测技术研究”(2019350000240137); 福建省烟草公司南平市公司科技计划项目“翠碧一号烟叶挂灰机理研究及其防控技术研究”（NYK2021-10-03）