Reasch On Manipulator Servo System Based On Binocular Vision

With the rapid development in recent years, Robot technology has become one of the importantdirection of the development of high-tech development, and it is mainly related to kinematics, sensortechnology, precision machinery, artificial intelligence and some other disciplines. Robotic arm havea extensive application in humanoid robots, service robots, or the field of industrial robots. Atpresent, most of Robotic arm are not autonomous obtain the information of the environment, insome applications, such as arc welding robot, handling robot, humanoid robot and space navigationfield of the robot, It need robotic arm obtain the information of the environment independently, andmake different actions based on the changes of the external environment. Manipulator visual servocontrol has a rapid development with the rapid development of machine vision technology, whichprovide the foundation for solving robot autonomous control in unknown environment.Visual servo control system include monocular servo control and binocular servo controlaccording to the number of cameras, and this thesis adopts binocular stereo vision technology.Based on the binocular vision measured position coordinates of the object in the space, so as tocontrol the robot arm to grasp of the object. Based on researching binocular stereo vision technologyand mechanical arm kinematics modeling, build a simple experimental platform for robot visualservo control. It mainly consists of image acquisition module, image processing of computer,mechanical arm and the mechanical arm controller. I mainly research vision system, stereocalibration, feature extraction, stereo matching and other related theories in binocular stereo visionof this paper, and according these theories, completed the algorithm design. The calibration ofbinocular vision system is used to calibrate the extrinsic and intrinsic of the cameras and the spacegeometric relationships; It can resume real camera model to the ideal model with the three-dimensional correction; The paper extract image features using the SURF feature point, andcomplete the stereo matching and remove mismatching based the SURF feature. Completekinematic analysis and kinematics of the manipulator, and complete the transformation in therelationship between the spaces coordinate of the end actuator and inertia manipulator of joints.Finally, achieve the debugging of the overall function of this system on the experimentplatform, verify the binocular stereo vision technology and mechanical arm motion control, and analyzed the accuracy of the system according to the experimental data.