In the past,machine vision technology has quickly stepped into industrial production,and its research and application in the industry have been increasing,especially in the field of automated production lines.Research on robot vision control processing production has followed closely.This paper mainly discusses the joint automatic cutting processing scheme which is mainly based on vision and supplemented by robot motion.In order to enable the robot to accurately and efficiently cut the surface of a cylindrical work-piece,a combination of vision reconstruction and robot motion technology is used to propose an algorithm for accurately generating a robot cutting trajectory based on binocular vision and laser target reconstruction points.Among the 3D reconstruction of part surfaces,panoramic tracking and extraction of trajectory curves has been a research hot-spot.The field of view of the monocular camera for visual reconstruction has limitations,and the cost of multi-vision stereo is too high.To this end,a dot laser transmitter is proposed to map the surface of the work-piece to form a target trajectory point as an involved solution.The process is as follows:(1)Set up a binocular CCD camera visual environment and supplement it with a dot laser measurement device.A dynamic trajectory is formed by traversing the surface of the work-piece,and then the pose rotation matrix theory and target two-way tracking algorithm are used.Finally,Halcon develops applications to achieve various attitude positions Mosaic and 360 ° panoramic target trajectory reconstruction and 3D coordinate extraction.(2)The coordinate of the target reconstruction trajectory point is desymbolized by NURBS to synthesize a continuous cutting processing path,and the efficient and convenient adaptive multiple discrete difference algorithm is used to plan the robot attitude trajectory.(3)Finally,through the Eye-to-Hand of the coordinate system of the robot vision hand-eye system,the conversion of the posture trajectory to the joint motion of the robot is solved,and a complete set of automatic processing schemes from 3D reconstruction curves to robot trajectory posture planning.The experimental data and simulation results show that the vision measurement system has high accuracy.In the reconstruction part,the registration error of the image target point is within 0.5piexl,and the binocular vision reconstruction error is below 0.4mm under the set camera calibration field;The stitching effect of each field of view trajectory points is very good,and its stitching accuracy is between 0.036 ~ 0.045 mm,ensuring that the panoramic three-dimensional coordinates of the work-piece contour can be accurately extracted;The robot's spatial curve attitude trajectory generated by multiple algorithms is real and effective,without defects such as flipping and abrupt changes.The attitude direction of its dominant cutting changes within 5 °,which meets the requirements for the continuity and smoothness of the attitude trajectory.The machining trajectory planned by the robot's hand-eye system has been verified by the simulation system and can complete the cutting task. |