This project was for the 2014 National Competition of Transport Science and Technology for Students in China. It aimed to extract real-time traffic flow parameters, including volume, density, and intersection delay, etc., based on videos captured by an unmanned aerial vehicle (UAV).
Video Capturing Equipment
The project used a quad-rotor unmanned aerial vehicle to capture real-time traffic flow video. Its flight height ranged from 60 to 80 meters.
A quad-rotor unmanned aerial vehicle used to capture real-time traffic flow video
Video Stabilization
A video stabilization method based on image-feature extraction and matching was developed to address the jitters produced by the UAV.
Demonstration of the video stabilization method
Vehicle detection and tracking
With stabilized video, the Gaussian mixture method was used to model the background, and foregrounds (moving vehicles) could thus be detected. A Kalman filter was then applied to track the detected vehicles.
Detection and tracking of moving vehicles
Traffic Flow Parameters Extraction
Based on vehicle detection and tracking, traffic flow parameters, such as flow volume, density, could be obtained. Moreover, by setting virtual loop detectors in lanes, arriving and departure moments could be recorded, which were then used to calculate intersection delays and service levels of the intersection.
Traffic parameter analysis platform based on unmanned aerial vehicle
Group Photo of the Winning Team
The project won the second prize of 2014 National Competition of Transport Science and Technology for Students in China. Here is the group photo of all the participants (2 teams) in Tongji University.
Group photo