• DocumentCode
    3294414
  • Title

    Occlusion adaptive object tracking based on video image and radar data

  • Author

    Juan Zhao ; Yao, Danya

  • Author_Institution
    Dept. of Autom., Tsinghua Univ., Beijing
  • fYear
    2006
  • fDate
    38869
  • Firstpage
    943
  • Lastpage
    946
  • Abstract
    Vehicle and pedestrian detection is important in traffic surveillance, and video image processing technology has evolved and offered promising prospects. Occlusion object tracking is difficult in image processing, and this paper gives a system using video image and radar data to segment the occlusion objects. In this system the objects are extracted by processing image sequence, and using the radar data for object tracking when there is an occlusion. First, in this paper, an improved method is proposed for background acquisition before object extraction and location, and then an algorithm is adopted for shadow clutter rejection. Second, object tracking with image and radar data is discussed. The experimental results show that this system can robustly track objects, and significantly improve the accuracy probability of tracking with occlusions
  • Keywords
    hidden feature removal; image sequences; object detection; radar tracking; video surveillance; background acquisition; image sequence; objects extraction; occlusion adaptive object tracking; pedestrian detection; probability; radar data; traffic surveillance; vehicle detection; video image processing technology; Clutter; Data mining; Image processing; Image segmentation; Image sequences; Radar detection; Radar imaging; Radar tracking; Surveillance; Vehicle detection; image processing; object tracking; occlusion; radar; sensor; traffic detection;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    ITS Telecommunications Proceedings, 2006 6th International Conference on
  • Conference_Location
    Chengdu
  • Print_ISBN
    0-7803-9587-5
  • Electronic_ISBN
    0-7803-9587-5
  • Type

    conf

  • DOI
    10.1109/ITST.2006.288709
  • Filename
    4068742