LI Hang,SHEN Jin-yuan,LIU Run-jie,et al.Tobacco Standard Library and Feature Selection Based on the Cluster[J].HEILONGJIANG AGRICULTURAL SCIENCES,2016,(04):140-143.[doi:10.11942/j.issn1002-2767.2016.04.0140]
基于聚类的烟叶标准库和特征选择
- Title:
- Tobacco Standard Library and Feature Selection Based on the Cluster
- 文章编号:
- 1002-2767（2016）04-0140-04
- Keywords:
- tobacco leaves; spectrum; clustering analysis; feature extraction
- 分类号:
- S572
- 文献标志码:
- A
- 摘要:
- 为实现烟叶自动化分级，采用聚类算法来剔除烟叶样本中异样样本，通过计算类间的方差，选取方差值大的特征作为有用特征。利用支持向量机模拟人工操作，进行13个等级分部位、分颜色的识别。结果表明：特征选择后的最佳准确率分别为95.52%、98.22%。在提升准确率的同时，减少了输入特征的个数和采集光谱数据所需的时间。
- Abstract:
- There are possibilities that the different samples may exist among the tobacco left samples,so the clustering method is applied to get rid of the different samples.By calculating the variance between the classes,the characteristics of big variance value as a useful feature were selected.Using SVM to simulate the artificial operation,like the recognition of the parts and the colors of thirteen grades of the tobacco left samples.The optima accuracy reached 95.52% and 98.22% respectively after the select of the useful features.Meanwhile,not only the accuracy arising,but the number of input features and the time needed for spectral data were reduced.
参考文献/References:
[1]
马文杰
.
烟叶图像采集技术规范与烤烟收购质量分级特征研究
[D].
武汉：华中农业大学，
2007.
[2]张惠民，韩力群，段正刚.基于图像特征的烟叶分级[J].武汉大学学报，2003,28(3):359-362. [3]Tatters filed G,Forbes K.Classification of tobacco leaves by color and plant position[C]//Proceedings of theNinth Annual Symposium of the Pattern Recognition Association of South Africa ,Cape Town,1998:11. [4]王毅，马翔，温亚东，等.应用近红外光谱分析不同年度工业分级烟叶的特性[J].光谱学与光谱分析，2012,32(11):3014-3018. [5]Hongmei Li,Jiande Wu,Kuake Huang,et al.Study of flue-cured tobacco classification model based on the PSO-SVM[J].Research Journal of Applied Sciences,Engineering and Technology,2013,5(19):4671-4676. [6]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. [7]彭丹青，申金媛，刘剑君，等.基于径向基网络的烟叶光谱分级[J].农机化研究，2009,53(10):15-18. [8]李航，赵海东，申金媛，等.基于BPSO和SVM的烟叶近红外有用特征光谱选择[J].物理实验，2015,35(6):8-12. [9]赵海东，申金媛，刘润杰，等.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):659-664.
相似文献/References:
[1]刘印锋.烤烟波谱特性的分析与识别[J].黑龙江农业科学,2010,(03):96.
LIU Yin-feng.Analysis and Identification of Flue-cured Tobacco’s Spectrum Characteristic of F;ue-cured Tobacco[J].HEILONGJIANG AGRICULTURAL SCIENCES,2010,(04):96.
[2]郭贺,金兰淑,林国林.利用近红外光谱法对烟叶氮钾含量的快速测定[J].黑龙江农业科学,2008,(04):103.[doi:10.11942/j.issn1002-2767.2008.04.0103]
GUO-He,JIN Lan-shu,LIN Guo-l in.Rapid Determination of Nitrogen and Potassium Contents in TobaccoLeaves by Near Infrared Reflectance Spectroscopy[J].HEILONGJIANG AGRICULTURAL SCIENCES,2008,(04):103.[doi:10.11942/j.issn1002-2767.2008.04.0103]
[3]王囡囡.追施普通尿素与控释尿素对玉米后期穗部叶片光谱特征的影响[J].黑龙江农业科学,2020,(07):44.[doi:10.11942/j.issn1002-2767.2020.07.0044]
WANG Nan-nan.Effects of Common Urea and Controlled-release Urea for Topdressing on Spectral Characteristics of Ear Leaves in Late Growth Stage of Maize [J].HEILONGJIANG AGRICULTURAL SCIENCES,2020,(04):44.[doi:10.11942/j.issn1002-2767.2020.07.0044]
备注/Memo
收稿日期：2016-01-28