Robotic Dynamic Target Recognition and Tracking Based on the Monocular Vision

Gather the information of the environment by the monocular vision. Using the H and S weight of the HSV color model, separate the target from the environment with a certain color, by a fast clustering algorithm for two-value image segmentation. Calculating the distance between the camera and target by the 3D reconstruction algorithm and sub-control strategy, and raise its veracity by laser information fusion. Furthermore, a vision servo system has been designed and utilized to achieve the robot's dynamic track. At last, some experiments were used to certification its availability.