Image processing based traffic monitoring in smart cities

OData support
Supervisor:
Dr. Vida Rolland
Department of Telecommunications and Media Informatics

This diploma work presents an application and its development. The purpose of the application is to create a traffic monitoring system using a concrete camera and its video stream. The application is designed to count and sort passing vehicles by their direction and size. In addition it detects local Volán buses passing in front of the camera. For image processing I used an open source function library, OpenCV. The program was written in Python.

After a short overview of different image processing methods, I present the most important OpenCV functions and especially those components that are useful for traffic monitoring. In the next section the components of the application are shown, including incurring problems, limitations and the possible solutions.

The last section presents the processing and the results. For this I used a time series database (InfluxDB) and an additional graphical displayer (Grafana).

As a result of my work, it can be stated that traffic counting and bus detection is possible with the use of a given camera, under certain conditions. However, while in daylight all functions of the program work properly, in darkness the sorting by size and the bus detection don’t reach the expected accuracy.

Downloads

Please sign in to download the files of this thesis.