[bULBot 02] Implementation and optimisation of real-time computer vision algorithms using a Zynq SoC
Each year, the bULBot club takes part in the Belgium robotic cup with one or two robots. The robots fights matches of 90s against other robots and any contact between robots must be avoided. To current system is not smart enough to identify the opponent if it is farther than 25cm.
The goal of this project is to implement computer vision algorithm to identify the location, velocity and direction of an opponent farther than 25cm using one camera (kinect) or several cameras.
To implement the needed image processing algorithms, a specific board with a Zynq SoC (dual core ARM + FPGA) and an Epiphany 16-core coprocessor will be used. Each processing ability of the board has its strengths and weaknesses and will be used for the part of the processing in which they are the most efficient.