Machine vision for a mobile robot

OData support
Kálmán Viktor
Department of Control Engineering and Information Technology

Recognizing of utilization of mobile devices and their increased spreading results in the need of their continuous developement. Nowadays these devices are used in solving of most varied tasks. It is clear that reliability, accuracy have key importance among the requirements against the appliance. It is necessary to ensure the robustness of systems, that requires accurate and deep understanding of working environment. The scope of this paper is the development and introduction of a visual sensing system, that is able to provide authentic information on the surroundings of the embedded device.

Beyond the general purpose (implementing an efficient vision system) the objective of the research team (me and my two colleagues who worked on different subsystems) was to prepare the platform to make success on the international EUROBOT competition. In order to accomplish the given task with the robot, it's indispensable to make coherence among the modules, given that their input and output data are strongly depend on each other. So that a reliable embedded system necessitates high level and elaborated communication.The steps of the concrete development was preceded by the analysis of solutions and methods created by students in the previous years. This served as a basic in solving the upcoming problems and difficulties, and eased the the transplantation of system-level approach to practice. The positioning objective is basicly dual, on the one hand we have to define the tierce-point of simple shapes and gleams in order to recognize them after the picture of the camera was taken, on the other hand we have to calibrate thus creating the relative direction required for positioning. The essay is closed by the presentation of system-level test results and appreciation.


Please sign in to download the files of this thesis.