Traking Detection Research Of Parallel Robot Visual Servo System

System of robot vision has widely area of research and range of applications which is an important area of machine vision research. In the robot visual servo system, the posture information of terminal manipulate to reflects the important parameters of robot movement state. According to these posture informations, we use of the robot geometric kinematics model for inverse solution of motion parameters. It has important significance to obtain the control parameters of the motion system. In the thesis, we use as research subjects which the Series GPM Redundant Parallel Manipulator two degrees of freedom, using visual tracking and detection methods, to establish visual detection platform and collect image of real-time the moving objects in the scene. Using industrial camera as vision transducer, to realize vision tracking and detection of the two degrees of freedom parallel robot, and integration and processing system of pose information and motion parameters.Including inside and outside parameters of CCD calibration, image acquisition and transmission, image preprocessing, planning motion path, visual tracking and detection, to build and analyze robot kinematics model, simulation trajectory of motion.We uses as a visual module which connected to Bumblebee2 industrial camera and image acquisition card. And then build a single-camera parallel robot based on visual model though study of the camera calibration technology. We calculate parameters matrix of the relative relationship which the camera coordinate system and space coordinates. It reduces the requirements of experimental conditions and simplifies process of the camera calibration between the complex robot system. We adopt fitting method of round base on Gate to tracking and detection target. The algorithm can achieves location of the centroid track, to reduces quantity of calculation target detection and increase speed of calculation. Gate tracking algorithm can tracks in real-time and accurate detection of the movement of parallel robot terminal manipulator which compares with the Kalman filter algorithms. Finally locates position and orientation parameters of two degrees of freedom parallel robot. It has smaller difference between the actually value can be used as feedback information to the robot controller's motion. In addition, we research of robot trajectory planning and control parameters related to the servo controller.The summary and vistas of the whole work are listed in the end.