Specification of the task

Bicycle usage is becoming more common lately. In order to develop the city infrastructure, which is appropriate for growing bicycle population, statistical information about number of cyclist in given part of the city is required.

Using video information from street crossings can be used to calculate the number of cyclist in each city part.

The goal of this project is to develop a method for bicyle counting from a video. The video should be taken at a street crossing.


The aim of this project is to detect bicycles at street crossings and use the obtained information for the improvement of the city’s infrastructure. Because of this, we used a static camera (such as the cameras used in the surveillance) and recorded many videos from different angles in some of the main squares in Graz.

The proposed solution would work on any video that is recorded using a stable camera that catches a bicycle from perpendicular point of view at least at one frame.

sources on GitHub


3 Members from 3 different cultures and Education background!

Abigél Mester

Mountain View

Nevena Krpež

Mountain View

Alexander Kolli

Mountain View