Computer Vision Traffic Sensor for Fixed and Pan-Tilt-Zoom Cameras

This project will develop and demonstrate the application of a traffic sensor based on computer vision technology for use on fixed and pan-tilt-zoom (PTZ) cameras for detecting and tracking vehicles, including motorcycles. Work in the first stage will involve installation and evaluation of two test bed systems in the states of New York and Illinois. For both systems, the hardware, height, power, and communication specifications will be developed. The systems will be installed with capability to remotely monitor and reconfigure each test bed from the home base in Clemson and their performance will be monitored in order to refine the detection and tracking algorithms and uploading new versions of the software as needed. Data will be collected continuously and downloaded at regular intervals, ensuring data availability for a variety of weather and environmental conditions. The system performance will be evaluated by comparing results with those obtained from existing sensors as well as with ground truth obtained by manually annotating the video data. This evaluation will include volume, speed, and vehicle classification. Work in second and final stage will involve installation and monitoring of two additional test beds at Traffic Management Centers in South Carolina. Logistical issues regarding network communication, camera sequencing, and PTZ operations will be addressed in collaboration with the DOT staff. The cameras to be monitored will have PTZ capabilities. Site visits to the camera locations will be made to measure camera parameters (including height and distance from the pavement) and to perform calibration to ensure that speed and classification data are accurate. The obtained results will be used to refine the calibration and tracking algorithms. Data will be collected in a variety of conditions and manually annotated in order to evaluate the performance of the system. The evaluation will include the accuracy of the calibration parameters themselves as well as the accuracy of the traffic parameters (speed, volume, and classification) obtained as a result of running the detection and tracking software on the calibrated system.

Language

  • English

Project

  • Status: Completed
  • Contract Numbers:

    Project 20-30, IDEA

  • Sponsor Organizations:

    Federal Highway Administration

    1200 New Jersey Avenue, SE
    Washington, DC  United States  20590

    American Association of State Highway & Transportation Officials (AASHTO)

    444 North Capitol Street, NW, Suite 225
    Washington, DC  United States  20001

    National Cooperative Highway Research Program

    Transportation Research Board
    500 Fifth Street, NW
    Washington, DC  United States  20001
  • Project Managers:

    Jawed, Inam

  • Start Date: 20090423
  • Expected Completion Date: 0
  • Actual Completion Date: 0
  • Source Data: RiP Project 22102

Subject/Index Terms

Filing Info

  • Accession Number: 01547245
  • Record Type: Research project
  • Source Agency: National Cooperative Highway Research Program
  • Contract Numbers: Project 20-30, IDEA
  • Files: TRB, RiP
  • Created Date: Dec 4 2014 1:00AM