Target recognition and localization of international standard platform humanoid robot

According to the rules of Chinese robot competition and RoboCup open standard platform league competition, the problem of the target recognition and localization with the NAO robot need to be solved. Through setting white balance of the robot's vision sensor, the images are processed and threshold's segmentation is got. Then the outline of the target can be got and the centroid is calculated. Thus the position of the target is located by structuring the geometric relation model between robot and targets.