基于聚类的烟叶近红外光谱有效特征的筛选方法

赵海东; 申金媛; 刘润杰; 刘剑君

红外技术, 2013, 35 (10): 659, 网络出版: 2013-11-01

基于聚类的烟叶近红外光谱有效特征的筛选方法

Tobacco Leaf Selection Method of the Near-infrared Spectroscopy Effective Feature Based on the Cluster

论文大纲

赵海东 ¹申金媛 ¹刘润杰 ¹刘剑君 ²

作者单位

¹ 郑州大学信息工程学院, 河南郑州 450001

² 郑州市烟草专卖局, 河南郑州 450006

近红外光谱特征选择烟叶分组聚类支持向量机 near-infrared spectrum feature select tobacco grouping cluster SVM

摘要

提出利用聚类方法对光谱数据进行特征筛选。通过分析类内参数 γ1和类间参数γ2对筛选结果的影响, 选择较好的γ1和γ2进行有用光谱特征筛选。利用烤烟叶的近红外反射光谱(1500～2400 nm间隔 2 nm), 选用 SVM方法进行部位和颜色分组识别, 训练样本的识别率为 100%, 测试样本的识别率分别是 96.22%和 92.79%。然后利用聚类方法对初始光谱进行特征筛选, 选用相同的 SVM方法及相同的学习样本和测试样本进行部位和颜色分组识别。在删减部分不相干光谱后, 识别率分别提高到97.23%和 95.52%;继续删除相关度不高的光谱, 在识别率略有下降时, 光谱特征数可减少到 200个以下。结果表明: 利用聚类方法进行特征筛选, 不仅可提高识别率, 且可大大减少光谱数据, 因而极大地减少了数据采集时间, 简化了分组模型, 提高了系统的实时和快速处理能力。

Abstract

The clustering method is applied to select the effective features from the original spectra. The effective features used for better γ1 and γ2 is chose by analyzing the influence of intra-class parameters γ1 and inter-class parameters γ2. Part and color of tobacco leaves are classified by SVM method based on the near infrared reflecting spectra(1500 nm-2400 nm interval of 2 nm)of flue-cured tobacco leaves. The recognition rates of part and color are 100% for train sample, 96.22% and 92.79% respectively for test sample. After some irrelevant spectra are removed by clustering algorithm, the recognition rate can be improved to 97.23% and 95.52% respectively. Continue cutting spectra having low correlation with classification. The recognition rate will declined significantly when too many spectra are removed. The number of spectra can be reduced to about 200 from 451 with slightly low recognition rate. The experiment results show that the clustering method can not only improve the recognition rate but also greatly reduce the number of spectral data. This greatly lessens the time of collecting data and significantly improves the real-time and fast processing ability of the system.

参考文献

[1] 常爱霞, 张建平, 杜咏, 等. 烤烟香型相关化学成分主导的不同产区烟叶聚类分析[J]. 中国烟草学报, 2010, 16(2): 14-19.

Chang Aixia, Zhang Jianping, Du Yong, et al. Cluster analysis of flue-cured tobacco leaves from different growing areas according to the chemical components correlating with aroma types[J]. Acta Tabacaria Sinica, 2010, 16(02): 14-19.

[2] 王维妙, 葛炯, 张建平. 近红外透射法预测再造烟叶中的5 种主要化学成分[J]. 烟草科技, 2009, 1(7): 43-47.

Wang Weimiao, Ge Jiong, Zhang Jianping. Prediction of five major chemical components in reconstituted tobacco sheet with NIR transmission spectroscopy [J]. Tobacco Science &Technology 2009, 1(07):43-47.

[3] 付秋娟, 王树声, 窦玉青, 等. 烟草根中N、K、Ca、Mg 的近红外光谱分析[J]. 烟草科技, 2006(10): 35-37.

Fu Qiujuan, Wang Shusheng, Du Yuqing, et al. Near-infrared analysis of nitrogen, potassium, calcium, and magnesium in tobacco root[J]. Tobacco Science &Technology, 2006(10): 35-37.

[4] 王东丹, 秦西云, 赵立红. 应用近红外光谱技术分析烟丝总糖和还原糖的研究[J]. 分析试验室, 2007, 26(5): 30-32.

Wang Dongdan, Qin Xiyun, Zhao Lihong. Study on analysis of total sugar and reducing sugar in cut tobacco by near-infrared spectrometry[J]. Chinese Journal of Analysis Laboratory, 2007, 26(5): 30-32.

[5] 杜阅光, 崔登科, 程小东, 等. 声光可调近红外光谱技术用于打叶复烤片烟化学成分[J]. 红外技术, 2012, 34(10): 614-618.

DU Yueguang, CUI Dengke, CHENG Xiaodong, et al. Homogenization self-control system of chemical composition in tobacco threshing and redrying process with acousto-optic tunable near-infrared spectroscopy[J]. Infrared Technology, 2012, 34(10): 614-618.

[6] 张建平, 陈江华, 束茹欣, 等. 近红外信息用于烟叶风格识别及卷烟配方研究的初步探索[J]. 中国烟草学报, 2007(5): 1-5.

Zhang Jianping, Chen Jianghua, Shu Ruxin, et al. Tobacco characteristics identification and blending formula study by using NIRs[J]. Acta Tabacaria Sinica, 2007(5): 1-5.

[7] 束茹欣, 王国东, 张建平, 等. 国产烤烟烟叶的NIRS 模式识别[J]. 烟草科技, 2006(8): 12-15.

Shu Ruxin, Wang Guodong, Zhang Jianping, et al. NIRS-based pattern recognition of domestic flue-cured tobacco[J]. Tobacco Science &Technology, 2006(8): 12-15.

[8] 王家俊, 汪帆, 马玲. SIMCA 分类法与PLS 算法结合近红外光谱应用于卷烟纸的质量控制[J]. 光谱学与光谱分析, 2006, 26(10): 1858-1862.

Wang Jiajun, Wang Fan, Ma Ling. The quality assessment of cigarette paper by SIMCA and PLS combined with near infrared spectrum[J]. Spectroscopy and Spectral Analysis, 2006, 26(10): 1858-1862.

[9] 章英, 贺立源. 基于近红外光谱的烤烟烟叶自动分组方法[J]. 农业工程学报, 2011, 27(4): 350-354.

Zhang Ying, He Liyuan. Auto-grouping method of flue-cured tobacco leaves based on near infrared spectra technology[J]. Transactions of the CSAE, 2011, 27(4): 350-354.

[10] 田高友, 袁洪福, 刘慧颖, 等. 小波变换在近红外光谱分析中的应用进展[J]. 光谱学与光谱分析, 2003, 23(6): 1111-1114.

Tian Gaoyou, Yuan Hongfu, Liu Huiying, et al. The application of wavelet transform in near infrared spectroscopy[J]. Spectroscopy and Spectral Analysis, 2003, 23(6): 1111-1114.

[11] 王立琦, 张礼勇, 朱秀超. 大豆油脂过氧化值的近红外光谱分析[J]. 食品科学, 2010, 31(6): 205-207.

Wang Liqi, Zhang Liyong, Zhu Xiuchao. Near-infrared spectroscopic analysis of peroxide value of soybean oil[J]. Food Science, 2010, 31(6): 205-207.

[12] 游伟, 李树涛, 谭明奎. 基于SVM-RFE-SFS 的基因选择方法[J]. 中国生物医学工程学报, 2010, 29(1): 93-99.

You Wei, Li Shutao, Tan Mingkui. Gene selection method based on SVM-RFE-SFS[J]. Chinese Journal of Biomedical Engineering, 2010, 29(1): 93-99.

[13] 郭文川, 王铭海, 岳绒. 基于近红外漫反射光谱的损伤猕猴桃早期识别[J]. 农业机械学报, 2013, 44(2): 142-146.

Guo Wenchuan, Wang Minghai, Yuerong. Early recognition of bruised kiwifruit based on near infrared diffuse reflectance spectroscopy[J]. Transaction of the Chinese Society for Agricultural Machinery, 2013, 44(2): 142-146.

[14] Araujo M C U, Saldanha T C B, Galvao R K H, et al. The successive projections algorithm for variable selection in spectroscopic multicomponent analysis[J]. Chemometrics and Intelligent Laboratory Systems, 2001, 57: 65-73.

[15] Shengfeng Ye, Dong Wang, Shungeng Min. Successive projections algorithm combined with uninformative variable elimination for spectral variable selection[J]. Chemometrics and Intelligent Laboratory Systems, 2008, 91: 194-199.

赵海东, 申金媛, 刘润杰, 刘剑君. 基于聚类的烟叶近红外光谱有效特征的筛选方法[J]. 红外技术, 2013, 35(10): 659. ZHAO Hai-dong, SHEN Jin-yuan, LIU Run-jie, LIU Jian-jun. Tobacco Leaf Selection Method of the Near-infrared Spectroscopy Effective Feature Based on the Cluster[J]. Infrared Technology, 2013, 35(10): 659.

基于聚类的烟叶近红外光谱有效特征的筛选方法

关于本站 Cookie 的使用提示

全站搜索

基于聚类的烟叶近红外光谱有效特征的筛选方法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索