• DocumentCode
    147641
  • Title

    Power enhanced extended maximum average correlation height filter for target detection

  • Author

    Bhuiyan, Sharif M. A. ; Khan, Jesmin F. ; Alam, Md Shamsul

  • Author_Institution
    Dept. of Electr. Eng., Tuskegee Univ., Tuskegee, AL, USA
  • fYear
    2014
  • fDate
    13-16 March 2014
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    Extended maximum average correlation height (EMACH) filter is generated from representative training images of expected target, and is designed to keep a balance between distortion tolerance and clutter rejection capability in target detection and tracking. However, in some cases it fails because of blurred target signature, similarity with background and clutter objects, and other distortions. Experimental study indicates that power enhancement of the images can overcome many instances of such failures. In this process, both the training images and test images undergo a power-law transformation, where the pixel intensities are raised to some power before developing the EMACH filter and/or performing the correlation operation. This process, indeed, high-lights the target image signature both in the filter and in the input scene and/or makes prominent difference between the target and other objects/backgrounds. The power enhanced EMACH filter, called PEMACH filter henceforth, has been found to be successful in several target detection problems.
  • Keywords
    image enhancement; object detection; target tracking; EMACH filter; blurred target signature; clutter rejection capability; image power enhancement; power enhanced extended maximum average correlation height filter; power-law transformation; representative training images; target detection problems; target image signature; target tracking; Clutter; Correlation; Integrated optics; Noise; Object detection; Training; Vectors; Extended maximum average correlation height (EMACH) filter; power enahanced EMACH (PEMACH) filter; power law transformation; target detection;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    SOUTHEASTCON 2014, IEEE
  • Conference_Location
    Lexington, KY
  • Type

    conf

  • DOI
    10.1109/SECON.2014.6950747
  • Filename
    6950747