• DocumentCode
    503925
  • Title

    Research on Fabric Image Registration Technique Based on Pattern Recognition Using Cross-Power Spectrum

  • Author

    Li Jiu-ling ; Wang Huang ; Zhao Da-Xing ; Zhu Jin-lei

  • Author_Institution
    Mech. Eng. Coll., Hubei Univ. of Technol., Wuhan, China
  • Volume
    2
  • fYear
    2009
  • fDate
    19-21 May 2009
  • Firstpage
    420
  • Lastpage
    423
  • Abstract
    This paper designs a fabric image registration method after researching texture of fabric. It can find the commonality area of two images quickly and accurately. Generally speaking, it is a good choice to solve the image registration with the relativity of two images in random texture images. However, repeated on high-resolution images to do Fourier Transform will decrease real-time of detection system. In order to enhance the real-time performance of fabric inspection system, the paper designs an effective method in the premise of ensuring matching effect. The method combines the excellence of frequency domain and space domain using predictable commonality area in the process of template matching, so arithmetic with favorable matching effect is put forward.
  • Keywords
    Fourier transforms; image matching; image recognition; image registration; image texture; object detection; Fourier Transform; cross-power spectrum; detection system; fabric image registration technique; fabric inspection system; high-resolution images; image matching; image texture; pattern recognition; Cameras; Design engineering; Fabrics; Fourier transforms; Image registration; Mechanical engineering; Pattern recognition; Pulp manufacturing; Real time systems; Software engineering; Fourier transform; cross-power spectrum; distributed software systems; image registration;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Software Engineering, 2009. WCSE '09. WRI World Congress on
  • Conference_Location
    Xiamen
  • Print_ISBN
    978-0-7695-3570-8
  • Type

    conf

  • DOI
    10.1109/WCSE.2009.117
  • Filename
    5319635