Ensure machine vision cameras are aligned to properly detect the onset of a queue.
Experience with advanced queue detection systems in Minnesota.
Made Public Date
08/30/2014

1063

Intersection of SH-55 and Rhode Island Ave
Minneapolis
Minnesota
United States

1064

Intersection SH-55 and Glenwood Ave
Minneapolis
Minnesota
United States

1065

WZ North of SH-280 and Broadway Ave
Minneapolis Metropolitan Area
Minnesota
United States
TwitterLinkedInFacebook
Identifier
2014-00681

Low-Cost Portable Video-Based Queue Detection for Work-Zone Safety

Background

This study tested the feasibility of using machine vision and portable traffic management systems (PTMS) to provide real-time detection of congestion at intersections and work zones, and inform drivers of proper speeds needed to safely approach growing queues. The machine vision system was enhanced with a control algorithm to provide real-time detection of queue onset, queue length, and trigger queue warning messages on dynamic message signs located upstream from the work zone control areas. The algorithm was designed to act like a trip-wire when video detection identified stopped vehicles within rapidly forming queues. The overall objective was to improve safety and prevent secondary crashes.

Although PTMS solutions have been widely deployed, systems using advance machine vision queue detection algorithms are still being explored. This study evaluated the technical performance of a prototype system and provided lessons learned.

Three sites in Minnesota were equipped. Two signal controlled intersection sites and one highway construction site. The algorithm was calibrated at each site and then video components of the system were used to evaluate performance.

Site 1 Description

The first site was an intersection located west of SH-100, at the crossing of Glenwood Avenue and SH-55. The camera was tilted down 20 degrees from horizontal, with a height of 28 feet above the roadway to provide a clear view of traffic well beyond 500 feet from the intersection.

Site 2 Description

The second site was an intersection at Rhode Island Avenue and SH-55. This site allowed for testing significant cross-lane view angles from an off-shoulder location covering three lanes of traffic. At this site, the camera was tilted down 30 degrees from horizontal, with a height of 28 feet above the roadway to provide a view of traffic just beyond 500 feet from the intersection.

Site 3 Description

The third site was located north of the Broadway Avenue Intersection along SH-280 at a work zone site with temporary traffic control measures in place. The taper zone was approximately 1000 feet. The system was deployed for 10 days at this location.

Lessons Learned

The following were recommended to ensure good system performance:

  • Ensure machine vision systems accurately detect the onset of a queue. Align the camera azimuth angle with the roadway as much as possible, with respect to the location of the stop bar, even if this necessitates a reduction in the queue length detection area.
  • Ensure the stop bar is visible. A stop bar must be completely visible and projected horizontally across all lanes to be monitored.
  • Consider the potential for cross-lane occlusions. Monitoring more than two lanes can be problematic if the system is deployed at a lateral distance of greater than one lane from the lane edge.

Low-Cost Portable Video-Based Queue Detection for Work-Zone Safety

Low-Cost Portable Video-Based Queue Detection for Work-Zone Safety
Publication Sort Date
01/01/2011
Author
Morris,Ted; Jory A. Schwach; Panos G. Michalopoulos
Publisher
University of Minnesota

(Our website has many links to other organizations. While we offer these electronic linkages for your convenience in accessing transportation-related information, please be aware that when you exit our website, the privacy and accessibility policies stated on our website may not be the same as that on other websites.)

Goal Areas

Focus Areas Taxonomy: