A MEASUREMENT SYSTEM FOR VEHICULAR QUEUE LENGTHS BY USING IMAGE PROCESSING

The authors have developed a measurement system for vehicular queue lengths in approach lanes of an intersection by using image processing. The data obtained in the system as information to control traffic signals allows one to estimate an optimal green time duration for releasing queuing vehicles from approach lanes before starting a new signal cycle. TV cameras for the system are installed at a height of 10 meters or so at an intersection to take images of traffic inflows at a bird's-eye view. In this system, a measurement road section is divided into unit intervals of an average of 30 meters long each. This system detects whether stopping vehicles exist in each unit interval or not on the basis of edge detection operations, and measures a vehicular queue length. This system using such a macroscopic detection method allows measuring queue length up to 300 meters long if the visibility is good. Since this system gives the threshold values to detect vehicles independently for each unit interval, it is able to measure queue lengths with high precision. To get the higher measurement precision, two TV cameras can be installed along a measurement road section; in this case, the image processing is done for the two divided parts as a continuous section. To verify the measurement precision of this system, the authors have experimented with actual traffic flow images in two road sections, 150 meters and 250 meters long. The images are taken in the morning, in the daytime including vehicular shadows, and at night. This system could measure vehicular queue lengths with a precision of not less than 95% regardless of weather environments.

• Supplemental Notes:
  • Full conference proceedings available only on CD.
• Corporate Authors:

  ITS America

  1100 17th Street, NW, 12th Floor
  Washington, DC  United States  20036
• Authors:
  • Iwasaki, Y
• Conference:
  • Intelligent Transportation: Realizing the Future. Abstracts of the Third World Congress on Intelligent Transport Systems
  • Location: Orlando, Florida
  • Date: 1996-10-14 to 1996-10-18
• Publication Date: 1996

Language

• English

Media Info

• Pagination: n.p.

Subject/Index Terms

• TRT Terms: Approach; Detectors; Estimating; Estimation theory; Green interval (Traffic signal cycle); Highway traffic control; Image processing; Intersections; Length; Measurement; Queuing; Systems; Television cameras; Time; Traffic lanes; Traffic signal timing; Traffic surveillance; Vehicle detectors; Vehicles
• Uncontrolled Terms: Edges
• Old TRIS Terms: Detection systems; Vehicular
• Subject Areas: Highways; Operations and Traffic Management; Vehicles and Equipment; I73: Traffic Control;

Filing Info

• Accession Number: 00741636
• Record Type: Publication
• Files: TRIS
• Created Date: Sep 23 1997 12:00AM