Development of a Robust Colour Classification Method for Parking Assist Systems

OData support
Supervisor:
Szemenyei Márton
Department of Control Engineering and Information Technology

Nowadays driver assistance systems gain more and more ground in the automotive industry. The cause for this is the race of corporations to satisfy market requirements, which is the lever for continuous improvements. Today the innovations are mainly influenced by customer needs, which are gounded in user comfort. The comfort of the driver is increased by better equipment that can help the owner concentrate fully on the driving, thus improving road safety. Since nowadays certain functionalities are demanded by everyone, they became part of the standard equipment, therefore increasing the customer’s attention is gradually getting harder. It follows that the interest of the customers can be best captured by features that are considered standard solutions, perhaps they do not know about the existence of them.

In my thesis I introduce an innovative solution that increases both the comfort, and accessibility of the vehicle. To improve traffic, the curb stones in numerous places can have various colours that have different meanings that, for example, in the USA can vary from state to state. My system is able to help the drivers with this problem, since after identifying the curb stone, it is able to perform colour classification and recognize the meaning of the color in the current location of the car.

In my thesis I develop this colour classification algorithm. I analyse currently available procedures and methods, after that I introduce the system design and develop my own solution. Using available sequences, I test my algorithm for robustness and runtime requirements, and I analyse it in a real environment. Finally, I evaluate the results and give suggestions for further improvements.

Downloads

Please sign in to download the files of this thesis.