Research And Design Of Hybrid Camera Visual Servo Robot System

To meet the production demands of the modern manufacturing industry,people have done on the robot and visual servo technology in-depth exploration.The visual servo technology provides rich control information for the robot in work,which makes it more flexible,and more adaptable to the production requirements under the complicated environment.On the modern industrial production line,visual servo system makes the industrial handling,assembly,welding,etc becoming more efficient,flexible and accurate.Therefore,in this paper,using SCARA robot as the research object,a hybrid camera visual servo robot system is established,and the detailed contents are as follows:Firstly,a velocity measurement method based on visual geometry is proposed.To simplify the visual geometry algorithm,the vanishing points and the vanishing lines on the image can be calculated off-line by fixing the position of Eye-to-hand camera and conveyor belt.The image processing methods of detecting feature points of workpiece mainly include image preprocessing,contour extraction and filling,ellipse fitting,Hough transform and so on.According to the extracted feature points,the actual displacements of the workpiece are calculated by using the visual geometry algorithm.In order to calculate the speed of workpiece moving accurately,this paper uses functions of the QueryPerformanceFrequency and QueryPerformanceCounter to record the time interval of workpiece moving.Finally,the kinematics formulas are used to calculate the moving speed of the workpiece,and the accuracy of the displacement and velocity of the workpiece are verified by experiments,respectively.Secondly,the vision system of mounting Eye-in-hand camera is calibrated.The internal parameters of the camera are calibrated by the MATLAB calibration toolbox.The rotation matrix of camera external parameters can be calibrated by controlling the robot gripper to do the orthogonal translation movement.A calibration method based on the laser pen is proposed for the calibration of the translation vector of camera external parameters.Through controlling the fourth connecting rod of the robot gripper mounted with a laser pen to do the rotation movement while maintaining the others connecting rods static,the feature points are determined,and according to the feature points,the translation vectors of camera external parameters are obtained.Finally,the parameters accuracy of the vision system calibration are verified by measuring the distance of known line segments in the scene.Thirdly,state estimation,grasping and precise placement of the moving workpiece.The target workpiece is identified and matched by calculating Hu invariant moments of the contours of workpiece and template which are extracted by the image processing algorithms.Then,position and orientation of the target workpiece are calculated by computing the centre of the workpiece image and the angle between the workpiece spindle and the robot coordinate system.In order to accurately place the workpiece to the target position,using the Eye-to-hand camera to capture the image of robot gripper grasping workpiece is proposed,and the actual deviation of centerline of robot gripper and the workpiece center which is obtained by the image processing algorithms is feeded back into the robot controller so that realizing the precise placement of workpiece.Fourthly,implementation of a hybrid camera visual servo robot system.The system realizes modular programming based on OpenCV library and the MFC class library provided by Visual C++ 2010.These include robot control module,speed detection module,workpiece positioning and grasping module,workpiece precise placement module and image processing module.Finally,the success rate of system grasping the workpiece and the accuracy of placing workpiece to goal position are verified by experiments,respectively.