Research On Visual Positioning And Grabbing Technology Of Industrial Robot

With the advent of "Made in China 2025",the degree of intelligence and customization in industrial production has continued to increase,especially the production line with industrial robots as the mainstay.With the empowerment of machine vision,it will show strong energy in response to the needs of multiple fields,multiple scenarios,and diversification in industrial production.The combination of machine vision technology and industrial robots has many advantages such as fast vision processing speed,high positioning accuracy,and strong versatility.It is widely used in assembly,palletizing,welding,spraying,cutting,polishing and other industrial production activities.In actual industrial applications,precise grasping of workpieces is the key research content of industrial robot technology.With the increasing demand for flexibility in the manufacturing industry,the use of machine vision technology to guide industrial robots to complete grasping has become a current research hotspot.This paper designs a visual positioning and grasping system for industrial robots in response to the needs of flexible production lines.The system takes the BRTIRUS0805 A industrial robot as the main body,and uses the bar light source,Real Sense D435 depth camera and Jetson TX2 image processor and other core hardware devices to form a vision subsystem.The system can automatically accomplish the task of locating and grasping randomly placed workpiece according to the visual information in the industrial application environment.The main research content of the subject is carried out through the following points:Firstly,the kinematics model of BRTIRUS0805 A industrial robot is established,and the mathematical model of the robot is established by using D-H theory.The forward and inverse kinematics formulas were derived by using numerical and geometric method respectively,and the forward and inverse kinematics formulas are simulated and verified by using MATLAB robot toolbox.Secondly,aiming at the problem of visual calibration.The principles and algorithms of camera calibration and hand-eye calibration are studied.The imaging model of the camera is established based on the pinhole imaging principle.A high precision camera calibration method was adopted and the results of camera calibration were optimized by L-M nonlinear least square method.The characteristics between the Eye-to-hand system and the Eye-in-hand system were compared and analyzed,and the hand eye calibration formula is derived.The Tsai hand eye calibration algorithm is used to solve the hand eye transformation matrix.Experiments are carried out for camera calibration and hand-eye calibration algorithms.The results show that the calibration results have high accuracy and can meet the needs of industrial applications.Thirdly,aiming at the problem of visual positioning.Various image processing algorithms are comparative studied.The image is preprocessed using the methods of integrated average gray scale,Gaussian filtering and threshold segmentation.The Canny edge detection algorithm and the Harris corner detection algorithm are used to extract features of the image.The geometric moment of the image is used to realize the precise positioning of the target workpiece,and the above algorithm is tested by using the actual workpiece images.Experimental results show that the algorithm has a good detection effect and can accurately locate the target workpiece.Finally,aiming at the problem of industrial application,using the algorithm principle of key technologies such as visual calibration and image processing,the software of visual positioning and grasping is developed.Combined with the hardware platform,the whole system is tested.The experimental results show that the absolute error mean and relative error mean of visual positioning are 2.484 mm and 0.54%,and the absolute error mean and relative error mean of robot's actual grasp are 3.102 mm and 0.56%.Compared with the traditional industrial robot,the proposed industrial robot visual positioning grasping system has less relative errors in the actual work process,and the system has higher working efficiency and intelligence,which can meet the needs of flexible production line.