VISION-BASED MONITORING OF INTERSECTIONS

This paper describes a passive vision-based sensing system to track vehicles and pedestrians in real-time to ultimately detect situations that lead to accidents. In order to work under different lighting conditions as well as clutter at the site, image segmentation is performed. Each pixel in the image is modeled as a mixture of Gaussian distributions. In this way the system learns the model of the background and updates it as lighting conditions change. The authors report the system has been used successfully using oriented bounding boxes in outdoor environments.

• Supplemental Notes:
  • Publication Date: 2002. IEEE Service Center, Piscataway NJ
• Corporate Authors:

  Guo li jiao tong da xue (China : Republic : 1949- )

  ,    

  University of Minnesota, Minneapolis

  Artificial Intelligence, Robotics, and Vision Lab
  Minneapolis, MN  United States  55455-0220

  University of Washington, Seattle

  Department of Electrical Engineering
  Seattle, WA  United States  98195

  University of Arizona, Tucson

  Center for Excellence in Advanced Traffic and Logistics Algorithms and Systems
  Tucson, AZ  United States  85721

  National Consortium for Remote Sensing in Transportation (U.S.)

  ,    

  National University of Singapore. Laboratories for Information Technology

  ,    

  Omron Corporation

  Traffic Solutions Division
  2-2-1, Nishikusatsu,
  Kusatsu-city, Shiga-prefecture  Japan  525-0035

  Ohio State University, Columbus

  Department of Electrical Engineering
  Columbus, OH  United States  43210
• Authors:
  • Veeraraghavan, Harini
  • Masoud, Osama
  • Papanikolopoulos, Nikolaos
• Conference:
  • IEEE Intelligent Transportation Systems Conference (5th : 2002 : Singapore). The IEEE 5th International Conference on Intelligent Transportation Systems : proceedings
• Publication Date: 2002

Language

• English

Media Info

• Pagination: p. 7-12

Subject/Index Terms

• TRT Terms: Arterial highways; Image processing; Traffic surveillance
• Subject Areas: Highways;

Filing Info

• Accession Number: 00974443
• Record Type: Publication
• Source Agency: UC Berkeley Transportation Library
• Files: PATH
• Created Date: Jun 2 2004 12:00AM