Classification in hyperspectral images by independent component analysis, segmentedcross-validation and uncertainty estimates
作者: Beatriz Galindo-PrietoFrank Westad
作者单位: 1Department of Engineering Cybernetics (ITK), Norwegian University of Science andTechnology (NTNU), Norway. beatriz.galindoprieto@ntnu.no
2Department of Engineering Cybernetics (ITK), Norwegian University of Science andTechnology (NTNU), Norway and CAMO Software, Oslo, Norway. frank.westad@ntnu.no
刊名: Journal of Spectral Imaging, 2018, Vol.7 , pp.1-11
来源数据库: IM Publications
DOI: 10.1255/jsi.2018.a4
关键词: Hyperspectral imagingROI selectionSpectroscopyIndependent component analysisICACross-validationUncertainty testJack-knifingHotelling’s T 2Classification
英文摘要: Independent component analysis combined with various strategies for cross-validation,uncertainty estimates by jack-knifing and critical Hotelling’s T 2 limitsestimation, proposed in this paper, is used for classification purposes in hyperspectral images. To the best of ourknowledge, the combined approach of methods used in this paper has not been previously applied tohyperspectral imaging analysis for interpretation and classification in the literature. The data analysis performedhere aims to distinguish between four different types of plastics, some of them containing brominated flameretardants, from their near infrared hyperspectral images. The results showed that the method approach usedhere can be successfully used for unsupervised classification. A comparison of validation...
原始语种摘要: Independent component analysis combined with various strategies for cross-validation,uncertainty estimates by jack-knifing and critical Hotelling’s T 2 limitsestimation, proposed in this paper, is used for classification purposes in hyperspectral images. To the best of ourknowledge, the combined approach of methods used in this paper has not been previously applied tohyperspectral imaging analysis for interpretation and classification in the literature. The data analysis performedhere aims to distinguish between four different types of plastics, some of them containing brominated flameretardants, from their near infrared hyperspectral images. The results showed that the method approach usedhere can be successfully used for unsupervised classification. A comparison of validation...
全文获取路径: PDF下载  IM Publications  (合作)
分享到:

×
关键词翻译
关键词翻译
  • component 成分
  • analysis 分析
  • validation 证实
  • uncertainty 不定
  • estimates 估值
  • knifing 刮膜
  • imaging 图像形成
  • combined 组合
  • paper 
  • brominated 溴化了的