Research On Robot Moving Workpiece System According To Demand Based On Visual Inspection

Grabbing and placing workpieces is one of the main tasks of robot systems in industrial environments.Visual control is an important control method for robot systems.The use of image technology to identify workpieces is of great significance in modern intelligent production lines.This paper mainly takes hinges as an example to carry out on-demand moving tasks,which can be used to replace manual workpiece placement before robot polishing,which helps to improve the automation and production efficiency of the robot polishing production line.This paper mainly studies the on-demand moving process of scattered and stacked workpieces.The main research contents are as follows:(1)Aiming at the difficulty in identifying the weak-textured stacked workpieces in the industrial field,an improved geometric template matching algorithm with holes on the workpiece surface is proposed,and the hinge is used as an example for workpiece identification.After completing the selection and calibration of the industrial camera,the Canny algorithm is first used for edge detection;secondly,the rotating jam algorithm is used to obtain the minimum area circumscribed rectangle of the contour.After geometric constraints,the contour of the hole corresponding to the hole is obtained,and random increments are used.The method calculates the minimum circumscribed circle of the hole contour to obtain the center coordinates of the hole feature;then adopts the proposed improved geometric template matching algorithm,that is,recognizes the workpiece according to the geometric constraints between the hole features;finally eliminates the significant edge of the workpiece according to whether the hole features exist Misidentification of artifacts.Finally,the improved geometric template matching algorithm based on geometric constraints proposed in this paper is used to perform workpiece recognition experiments on hinges,L-shaped connectors,T-shaped connectors and flanges,which proves the effectiveness of the method.(2)Design and build a robot moving workpiece system on demand.First,build a virtual debugging platform for robots to move workpieces on demand based on the Siemens NX MCD platform to verify the rationality of the layout and study the path planning of the robots to move workpieces on demand,laying the foundation for the construction of a real robot on-demand workpiece moving system.According to the functional requirements of the workpiece ondemand moving system,design the system scheme,build the robot on-demand moving workpiece system,and design the robot control system,end buffer device and host computer interface,and finally describe the pose of the rigid body and the conversion of the coordinate system The explanation is convenient for subsequent robot pose conversion and robot motion control.(3)Carry out experiments on robots moving workpieces on demand.To perform robot grasping,accurate pose information of the workpiece is required.First,the pose estimation of the workpiece is performed.Since the actual size of the workpiece is known and the position of the hole feature in the image has been obtained during the workpiece recognition stage,the Pn P problem is solved Methods The pose estimation of the workpiece is carried out.The P3 P method,EPn P method and the iterative method-based pose estimation results that are commonly used in solving Pn P problems are compared and analyzed through experiments.Finally,the iterative method with smaller reprojection error is selected as the robot to move the workpiece on demand The pose estimation method of the system.On the basis of the hand-eye calibration of the robot's on-demand moving workpiece system,the coordinate transformation of the workpiece is performed to obtain the pose of the workpiece in the robot base coordinate system;then based on the path planned in the process of moving the workpiece on demand,the robot is sport control.Finally,a workpiece moving experiment was carried out on the built workpiece moving platform,which verified the effectiveness of the method proposed in this paper.This paper proposes an improved geometric template matching algorithm characterized by holes on the workpiece surface,which solves the problem of difficult identification of stacked workpieces with holes and weak textures in the industrial field;proposes a virtual debugging method based on NX MCD,and the layout of the virtual debugging platform is reasonable The research on path planning of the process of moving parts and moving parts shortens the design cycle and reduces costs;the iterative method is selected as the method for estimating the position and pose of the workpiece with the relative position relationship between the known hole features,and then the hand-eye calibration and the position and pose conversion are performed to obtain the accurate workpiece Posture information,and finally the workpiece moving experiment.It provides an effective method for robots to move workpieces on demand,which helps to improve the automation and production efficiency of the robot grinding production line,and increase economic benefits.