A VEHICLE RECOGNITION METHOD ROBUST AGAINST VEHICLES' OVERLAPPING BASED ON STEREO VISION

This paper presents a stereo-based method for recognizing overlapped vehicles. The method has three key features: 1) feature extraction based on a size-changeable feature extraction window, 2) camera calibration using environmental information, and 3) model matching with various overlapped vehicles' models. Experimental results reveal an average recognition rate of above 95% for images that which have overlapped vehicles.

• Supplemental Notes:
  • Publication Date: 1999. IEEE Service Center, Piscataway NJ
• Corporate Authors:

  Nippon Shingo Kabushiki Kaisha

  ,    

  Ruhr-Universitat Bochum. Institut fur Neuroinformatik

  ,    

  Universitat-Gesamthochschule-Siegen

  ,    

  Hitachi Seisakujo

  ,    

  Osaka Daigaku

  ,    

  Keisatsu Daigakko

  ,    
• Authors:
  • Kimachi, Masatoshi
  • Wu, Yue
  • Ogata, Shiro
• Conference:
  • IEEE/IEEJ/JSAI International Conference on Intelligent Transportation Systems (1999 : Tokyo, Japan). Proceedings : 1999 IEEE/IEEJ/JSAI International Conference on Intelligent Transportation Systems
• Publication Date: 1999

Language

• English

Media Info

• Pagination: p. 865-869

Subject/Index Terms

• TRT Terms: Cameras; In vehicle sensors; Pattern recognition systems

Filing Info

• Accession Number: 00922608
• Record Type: Publication
• Source Agency: UC Berkeley Transportation Library
• Files: PATH
• Created Date: Apr 30 2002 12:00AM