From jderobot
Jump to: navigation, search

This project deals with an application that it is able to carry out precise estimation of vehicle speeds. Besides, it counts the number of the cars that pass on the road. For this reason, we have used one single camera. The camera is the only source of information. The images it provides are analyzed to give estimations of vehicle speeds. The CarSpeed application detects and tracks the vehicles on the road. The detection stage allows the system to note the new vehicles coming into the image. The tracking stage starts when the system detects a vehicle, and uses an evolutive algorithm. Along these two stages the system deals with the position of the cars in the plain road, we have used an homography between road plane and the image plane to get absolute positions for the vehicles.


Redouane Kachach (redo.robot AT, JoseMaria Cañas (jmplaza AT and Victor Hidalgo (hbmhidalgo AT share the credits of this application.

Related links:


  • Car classifier
  • Results with real cars (A5 Highway at Móstoles), outgoing from the camera location.
  • Results with real cars (Highway at Villaviciosa de Odón), ingoing to the camera location.

Classification example on a real highway (A6 Madrid)

 In the following videos we can see some examples of the classification process on a multi-lane (Madrid, A6 highway). The following
 colors are used to distinguish between the different classes:
  • Motorcycles: Yellow
  • Cars: Orange
  • Vans: Blue
  • Trucks: Pink
  • Road-Train: Green
 Each tracked/classified vehicle is assigned a unique number during the classification process. At the end of the
 tracking/classification zone at the right of the video you can see the final assigned class and the estimated speed 
 in km/h for this vehicle. The following abbreviations are used for the different classes:
  • Motorcycles: MA
  • Cars: CA
  • Vans: VA
  • Trucks: TR
  • Road-Train: RD