The objective of this project was to conduct “proof of concept” evaluation to document the potential for Video Analytics as a tool for traffic operations centers (TOCs) and for traffic data collection. The project reported results from testing of several Video Analytics Systems in the United States (Iowa, Missouri) and in Ontario, Canada, under real-world environments. Components of the project evaluation included:
- Traffic Data Collection - Compared Video Analytics outputs to trusted agency detectors, for traffic volumes, speeds, and vehicle classifications.
- Incident Detection – Compared Video Analytics alerts to still images, video clips, and agency-reported incidents, to validate accuracy for detection of incidents including stopped vehicles, debris in the road, and slow traffic/congestion.
- Wrong-Way Vehicle Controlled Test – Conducted a controlled test to test the ability of Video Analytics systems to detect wrong-way movements on freeway ramps.
- Observations from Agency Staff: Representatives from participating agencies provided feedback on the value and usefulness of the Video Analytics systems to assist with managing road networks.
Results indicate that Video Analytics is ready to support a number of operational uses in its current state of practice. The final report includes several use case scenarios, along with procurement guidance and readiness of the state of practice to support each scenario tested as a part of this project. Lastly, a number of lessons learned are documented, which agencies can use as they plan for and procure Video Analytics systems.
The Next Generation Traffic Data and Incident Detection from Video Project was presented at the following:
- ITS Canada 2014 Annual Conference and General Meeting (June 2014)
- 2014 National Rural ITS Conference (August 2014)
- 2014 ITS World Congress (September 2014)
- ITS Minnesota 2014 Fall Forum (October 2014)
- ITS Midwest 2014 Annual Meeting (October 2014)
- ENTERPRISE Project Webinar (November 2014)