Position And Pose Detection For 3-PRS Parallel Mechanism Based On Binocular Vision

The parallel robot/mechanism has complicated structure,and has widely obtained attention with the advantages of powerful carrying capacity,small accumulative error,high precision and so on.The research of Lower-DOF parallel robots with three degree of freedom has become a hot research topic in recent years.The terminal position and pose are important parameters reflecting the motion state of the parallel robot,and they are also important indexes in judging work performance of the parallel robot.So it is very important to measure the terminal position and pose of the parallel robot accurately.Compared with the traditional contact measurement,binocular vision has no effect on contact force in measurement and has the advantages of high intelligence,fast detection speed,high accuracy and non-destructive measurement and so on.The binocular vision measurement is very effective for the detection of the position and pose of the parallel mechanism with multiple degrees of freedom and complex trajectory.The 3-PRS parallel mechanism is used as the detection object to study the binocular vision measurement in this paper.The main contents are as follows:(1)Kinematics analysis of parallel mechanism.According to the structure characteristics of the 3-PRS parallel mechanism,the kinematics model of parallel mechanism is established.The inverse kinematics and positive kinematics are analyzed.The position and pose equation of the parallel mechanism is deduced.The error between the position and pose calculated by the positive solution and the given position and pose parameters is small.The correctness of the kinematic model is verified by this example.(2)The overall Scheme's design and analysis for position and pose detection of 3-PRS parallel mechanism.The overall scheme for position and pose detection of 3-PRS parallel mechanism is designed.The binocular vision detection model is established,and the relationship between the three-dimensional coordinates and the image points of any spatial point is analyzed and the visual measurement equation is obtained.The left and right cameras in the detection system are calibrated respectively and their internal and external parameters are obtained.Then the stereo calibration is carried out and the geometric relationship between the two cameras is further analyzed.(3)Setting,identifying and locating the mark points of parallel mechanism.The mark points to reflect the terminal position and pose are established.The collected images are filtered and enhanced to realize the recognition and positioning of the mark points.The mark points in the two camera images are stereo-matched and the three-dimensional reconstruction of the spatial points in the two cameras.Finally,the detection of the terminal position and pose of the moving platform is realized.(4)The experiment and analysis of the position and pose detection.The experimental platform is built,the components of the hardware platform are selected,and the hardware platform is built.In order to observe the movement of parallel mechanism visually,the software is used to design human-machine interface.The experimental results are analyzed to verify the correctness and effectiveness of the proposed method.