Research On Calibration Of Binocular Active Visual Monitoring Platform Of Parallel Robots And Its Motion Control

The machine vision is to replace human eyes with machines in terms of measuring and judging, which achieves advanced technology of the system intelligence, automation and information by means of the vision sense technology, image understanding technology, pattern recognition and computer control technology. In order to improve the active recognizing ability of the parallel robot, the vision technology combine with the parallel robot technology to build a binocular active visual monitoring platform of parallel robot (BAVMPOPR) based on circular guide rail. It has greatly theoretical and technical value for improving the movement accuracy of parallel robot and facilitating the practicality of parallel robot. This paper does research on the calibration system and motion control under the BAVMPOPR which has been accomplished.Firstly, the introduction of BAVMPOPR that two cameras can be flexible movement along the circular orbit has been put forward, and the correlative theories of camera vision model and calibration method referred to BAVMPOPR are described.Secondly, according to the prototype of BAVMPOPR, Using the method of matrix transformation establish its kinematic model. The solving methods of forward and reverse position, control pulse equivalent and structure parameters are given.Thirdly, aiming at the problem of the dynamic calibration of the cameras in the BAVMPOPR, a calibration system is designed to install six groups of 3D target blocks which are uniformly distributed in a circular orbit inside.Finally, on the base of kinematic model and calibration system, combing with the constraints of kinetic control model, current position of moving target and optimal observation mode. The visual tracking model of binocular emmetropia mode is established with the goal of its continuity and minimum cost.