Goal Detection and Opponent Avoidance Algorithm for Wheeled Robot Soccer using Color Filtering and Contour Extraction

Abstract

This paper proposed a goal detection and opponent avoidance algorithm for robot soccer using color filtering and contour extraction. The paper presented a robust and simple algorithm based on an image captured by robot soccer to recognizing and locating a goal and avoiding from the opponent. The proposed algorithm used a color filtering in the HSV color space to distinguish and detect goal from the background image. The detected goal was then extracted using contour extraction to ensure the detection result. To reduce the noise, the morphology filter was applied to the detected object. The opponent robot was also detected by using HSV color filtering. The opponent robot has a designation color team that is different with our robot team. In order avoiding from opponent robot after detecting its color, we used ping sensor to detect the distance between robots. In the certain distance, our robot could avoid collision with an opponent robot with moving to the left or right side of the opponent robot. The experimental result showed the satisfactory result in detecting goal and avoiding an opponent. The goal successfully was detected from 1.5 m until 8.5 m distance from our robot. The robot successfully avoided the opponent robot in the distance of 40 cm, The robot moved to the right, or left side depended upon the opponent position.