• DocumentCode
    2113547
  • Title

    Design of Comprehensive Color Feature Vector for Tile Color Online Inspection

  • Author

    Li, Weiqing ; Wang, Chengbiao ; Wang, Qun ; Chen, Guangshe

  • Author_Institution
    Sch. of Eng. & Technol., China Univ. of Geosci., Beijing
  • Volume
    2
  • fYear
    2008
  • fDate
    20-22 Dec. 2008
  • Firstpage
    24
  • Lastpage
    27
  • Abstract
    This paper proposed a comprehensive color feature vector for tile online color consistency inspection. Non-uniform color quantitative method that was based on main color and human vision was put forward and the result accord with the human vision and the speed meet the requirement of tile online inspection. In order to synchronously meet the demand of tile global and local color difference inspection, especially to the tile with texture, global and composite color center of main color were introduced to describe tile color spatial distribution information. A comprehensive color feature vector including tile color global statistical information and spatial distribution information was designed and solved the difficult problem of tile color difference inspection and classification. The results show that this method was suitable for not only monochrome tile but also the tile with complex texture.
  • Keywords
    image classification; image colour analysis; image texture; inspection; statistical distributions; color classification; color difference inspection; comprehensive color feature vector design; global statistical information; image texture; nonuniform color quantitative method; spatial distribution information; tile online color consistency inspection; Color spatial distribution; Non-uniform color quantization; color feature;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Science and Engineering, 2008. ISISE '08. International Symposium on
  • Conference_Location
    Shanghai
  • Print_ISBN
    978-1-4244-2727-4
  • Type

    conf

  • DOI
    10.1109/ISISE.2008.312
  • Filename
    4732335