Colored Ball Position Tracking Method for Goalkeeper Humanoid Robot Soccer

Abstract

The goalkeeper robot in a robot soccer game must be able to detect position of the ball and block the ball into the goal. Identifying the location of the ball using robot vision becomes very important in addition to the mechanical motion of the robot. This paper proposed method to track colored ball position for goalkeeper robot. First, image is captured with web camera and then color filtering based on HSL color model performed on each image frame in the video. After that, images from filtering process are detected its blob by labeling nearest connected components. The largest blob in image represents the ball. The next step is detection of ball position using 9-Cells coordinate. Ball position coordinate is updated in some period of time and then sends to robot controller to conduct the robot's movements appropriately. Experiment results show that system is able to detect the ball and its position in 9-Cells coordinate which can be used by goalkeeper robot soccer controller to block the ball.