Traffic Monitoring By Video: Vehicles Tracking and Vehicle Data Analysing

CS 585 Final Project
Wenda Qin
Tian Zhang, Junhe Chen
Date: 12/5/16

Problem Definition

In this project, we use techniques of computer vision to find out vehicles on videos that captured from common roads, track the vehicles' traces and calculate the velocity and numbers of the cars in the video.

Method and Implementation

We use Harris Corner Detector to find out cars in the video, then, we use Kalman filter and Hungarian algorithm to track the traces of each car, and finally
use the knowledge of camera calibration to help calculating the velocity of our tracked vehicles

the program consists of four parts: main, tracker corner detection and helper. For the main part, we use it to combine other three parts to implement the main flow of our program.
For the tracker part, it contains three parts: Ctracker;HungarianAlg;Kalman. We use tracker part to establish a tracker to track traces of cars founded by corner detection

and use Kalman filter and Hungarian Algorithm to finish its prediction and matching

For the helper part, it mainly contains encapsulated functions to get or show the our tracking result, such as getting last position of points in a trace, drawing lines of traces in the target image
Briefly outline the functions you created in your code to carry out your algorithmic steps described above.



Processing Steps
Original Step
Origin video
Corner Dectection
Next step
Kalman Filter
Speed Calculation


By using corner detection, we can distinguish cars on our image and effectively avoid some occlusion and lighting problems. By using Kalman filter and Hungarian algorithm, we can detect a large part of cars that move across the image. However, these two methods will also bring errors respectively due to their limitation, we need to do some modifications on them so as to lower their negative influences.

Credits and Bibliography

[1]Benjamin Coifman, David Beymer, Philip McLauchlan, and Jitendra Malik. A real-time computer vision system for vehicle tracking and traffic surveillance. Transportation Research Part C: Emerging Technologies, 6(4):271 – 288, 1998.
[2]Bas and Crisman, 1997 Bas, E.K., & Crisman, J.D. (1997). An easy to install camera calibration for traffic monitoring. Proceedings of IEEE Conference on Intelligent Transportation System, Boston, Massachusetts,.(pp. 362-366).

[3]Lütteke, Felix, Xu Zhang, and Jörg Franke. "Implementation of the hungarian method for object tracking on a camera monitored transportation system." Robotics; Proceedings of ROBOTIK 2012; 7th German Conference on. VDE, 2012.

[4]J. Munkres, "Algorithms for the Assignment and Transportation Problems," Journal of the Society for Industrial and Applied Mathematics, Vol. 5, No. 1, pp. 32-38, 1957.

Code References: MultipleObjectTracking: