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
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
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.

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.

本文已被 2 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!