The traffic in our roads is significant and most part of it cames from road transportation and personal vehicles. The peak hours and the number of vehicles using the road are one of the most valueable information a bigger city or a highly used road has. Such information can be used to plan road network development and the to understand the flow of traffic in peak hours and reduce congestion. Traffic counting have been done in the past using various techniques but all of them required a lot of manpower. The aim of this project is to change that alltogether by removing the human factor as much as possible while counting vehicles based on a video.
The program has two main components. One part, the algorithm is responsible for couting the vehicles and providing information about them if it given the correct configuration. The other part is the graphical user interface which supports the tracking algorithm by providing the configuration and giving the opportunity to set them up and displaying the tracking’s results.
The software is capable of doing a high precision tracking if the project it’s working on is set properly. This means accurate tracking of the number of the vehicles moving through as well as their direction and being able to display those data multiple ways. After the preparations have been done it is possible to run automated trackings which need no human interaction.
The tracking can be done using one of the two methods at this time. One of them is a more robust but slower algorithm and the other one is a way faster but more special. The faster one cannot be used so widely due to the more special conditions it needs. The aim of the further development is to choose the better one or to make it possible to use both working together to achieve better results.