A real time method to object detection and tracking applied to robot-soccer

This paper proposes a computer vision system to a robot-soccer team, which specifications follow the stated by MiroSot (Micro-Robot Soccer Tournament) league, organized by FIRA (Federation of International Robot-Soccer Association). This system consists in a calibration module followed by a robust real-time computer vision module applied to detect and track the objects in the playfield. The very fast processing time and the robustness under smooth light variations are the major contributions of this paper.