Research On The Technology Of Grasping Large Size Flexible Workpiece Based On Machine Vision

Industrial robot has the characteristics of high efficiency,flexibility and accuracy,which is suitable for grabbing workpiece on assembly line.For large size flexible workpiece with random position,easy deformation and large size error,How to position the workpiece and adjust the shape of the workpiece,so that the robot can grasp the workpiece accurately,is the key to make the robot more intelligent.The machine vision technology is applied to the robot grasping process,which makes the robot have the function of human eyes.It is of great significance to guide robot to grasp workpiece through machine vision.In this paper,taking the actual demand of sponge mattress production as the starting point,the grasping process of large-scale flexible workpiece based on machine vision is studied,and a robot grasping system with visual information is constructed.The main research contents are as follows:(1)According to the characteristics of the workpiece,the hardware of the system is selected.In view of the lack of the current market research on large-scale lighting technology,a set of light source system is designed to meet the lighting needs of large-scale workpiece,which ensures the quality of the acquired image;(2)A machine vision guidance system is designed,and the mathematical model among the coordinate systems of robot,camera and workpiece is established.The distortion of the camera is eliminated by using the calibration method of Zhang Zhengyou,and the problem of "seeing accurately" is solved.The transformation relationship between hand eye transformation matrix and workpiece height is deduced by pinhole camera model,which provides a mathematical support for the robot to determine the grasping position and posture of the workpiece.(3)The positioning algorithm based on corner points and the correction algorithm based on contour difference are studied.In view of the random position of the large-scale workpiece,the center position and rotation angle of the workpiece are calculated by the improved Hough transformation,which realizes the automatic positioning of the workpiece and provides the position and posture of the target point for the grasping robot.In view of the problem that the flexible workpiece is easy to deform and the size cutting is not accurate,which leads to the poor composite accuracy of the mattress,an adaptive correction algorithm based on the contour difference of the workpiece is proposed.The method realizes the automatic rectification in the process of grasping the workpiece and ensures the product accuracy.(4)Based on the research of the key vision technology of positioning and rectification,a large-scale flexible workpiece grabbing system is designed,and the validity of the algorithm of workpiece positioning and rectification based on machine vision is tested,and the actual grabbing test is carried out.The test results show that the average working time beat of the system is less than 20 s,and the composite precision is within 2 mm,which meets the actual production requirements.