• DocumentCode
    400038
  • Title

    Stabilizing image sequences taken by the camcorder mounted on a moving vehicle

  • Author

    Liang, Yu-Ming ; Tyan, Hsiao-Rong ; Liao, Hong-Yuan Mark ; Chen, Sei- Wang

  • Author_Institution
    Inst. of Inf. Sci., Acad. Sinica, Taipei, Taiwan
  • Volume
    1
  • fYear
    2003
  • fDate
    12-15 Oct. 2003
  • Firstpage
    90
  • Abstract
    This paper presents an image stabilization technique which can be used as a very useful on-board unit of a vehicle. The proposed technique includes three steps. First, we detect two lane markings and their corresponding vanishing point from the input image. These two lane markings and their corresponding vanishing point are then used as the bases to calculate the camera motion parameters. The proposed motion estimation process can be used to solve the difficult problems such as feature point extraction and correspondence matching. In addition, the extracted road information can be used to tackle more advanced traffic problems. In the last step of our approach, a motion smoothing process is developed to remove some unwanted motion. This step is fairly important for obtaining very stable image sequences. The experimental results have shown that our method is indeed superb for stabilizing image sequences.
  • Keywords
    feature extraction; image matching; image sequences; motion estimation; road vehicles; stability; video cameras; camcorders; camera motion parameters; feature point extraction; image matching; image sequences; image stabilization technique; lane marking detection; motion estimation process; motion smoothing process; moving vehicles; Automotive engineering; Cameras; Feature extraction; Image sequences; Motion detection; Motion estimation; Roads; Smoothing methods; Vehicles; Video equipment;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE
  • Print_ISBN
    0-7803-8125-4
  • Type

    conf

  • DOI
    10.1109/ITSC.2003.1251927
  • Filename
    1251927