INDOT was honored for being one of the Sweet 16 selections among the annual High Value Research Project submissions at this summer’s American Association of State Highway and Transportation Officials (AASHTO) Research Advisory Committee annual meeting in Chattanooga, Tenn.
The national competition emphasizes the benefits of research and implementation strategies by state departments of transportation. The Sweet 16 consists of the top four identified projects from each of the four AASHTO regions. More than 200 projects were nominated nationwide.
INDOT’s research project — “Road Condition Detection and Classification from Existing Closed-Circuit Television (CCTV) Feed” — was done in conjunction with Purdue University’s Joint Transportation Research Program (JTRP).
Research & Development (R&D) Manager Tim Wells and Intelligent Transportation Systems Engineering Director Ed Cox played crucial roles in the project.
INDOT has more than 500 digital cameras to monitor traffic and incidents along our roadways in populated areas. The videos from the cameras are observed one by one by human operators. The main objective of the project was to develop an automatic, real-time system to monitor traffic conditions and detect incidents automatically based on INDOT’s CCTV feed.
The team designed hardware and software components to add to the existing CCTV system, the database structure for traffic data extracted from the videos, and a user-friendly web-based server for showing the incident locations automatically. Specific work in the project included detection of traffic-flow rate, traffic conditions, and traffic incidents, as well as classification of vehicles involved in incidents.
The system used many personal computers. Each camera was connected to a PC, which ran an artificial intelligence-based video processing program to monitor each highway lane in real time. All PCs uploaded the lane information to the cloud database and displayed it on a website that human operators accessed from the INDOT network.
“There are video analytics all over the country, and analytics are easily pulled from fixed cameras,” said INDOT Transportation Systems Management Operations Director Jim Sturdevant. “But everybody in the transportation industry uses pan-tilt-zoom cameras. The
research team was able to make the video analytics work on pan-tilt-cameras, and that’s why it’s so high value. Kudos to them.”