Development And Application Of Industrial Robot Deburring Platform Based On Monocular Vision

In order to adapt to the new national defense and military industry,automobile manufacturing,semiconductor production and other fields,industrial robots,which emerged in the 1950s,urgently need to integrate traditional manufacturing technology with modern information technology.In the modern production workshop,the machine vision system is added to the robot body to give the machine arm vision function.Through the "hand-eye" cooperative work,it can automatically complete the welding,handling,sorting,processing and other tasks of all kinds of workpieces,greatly improving the production efficiency of enterprises.At present,for small batch,different shapes,different sizes of small workpiece identification detection,deburring applications,most use binocular vision positioning method,this method simulates the human binocular vision three-dimensional imaging process,high requirements for hardware precision and equipment installation,background image processing algorithm is time-consuming and complex,can not be flexibly built according to the industrial site.In this paper,an industrial robot hand eye collaborative application platform based on monocular vision is developed on the laboratory industrial robot platform.The platform can transfer the pose information and edge contour information of the workpiece to the robot controller,which guides the robot end effector to remove the edge burr of the workpiece,Finally,the research and experiment of the corresponding vision system and robot system are completed to verify the accuracy of the hand eye cooperation system.This paper will study from the following four aspects:Considering the structure size and relative installation position of each equipment(machine vision device can not affect the free movement of robot arm),compatibility(visual processing software and robot controller can be interconnected),a deburring platform of industrial robot with machine vision device is built,and a specific light source device is designed to reduce the reflected light entering the lens.The camera is fixed outside the body of the robot,and the eye to hand method is used to recognize the two-dimensional space layer.Combined with Halcon software and C#language,the corresponding upper computer software is developed and designed;The relative spatial position relationship among the end effector of industrial robot,industrial camera and workpiece to be identified is analyzed.Through two steps of camera calibration and hand eye calibration,the pixel level size is converted to the actual physical level size.Finally,the accuracy of hand eye calibration is verified by measuring the workpiece with known size by vision;The preprocessing of the original image includes two steps:eliminating the surface noise and enhancing the edge of the workpiece;The region of interest(ROI)is defined to reduce the image processing area.After comparing different edge detection methods,Canny operator is selected to detect the contour of each edge object,and the workpiece template is extracted based on similarity template matching method.After successful matching,a complete and continuous set of pixel edge points is extracted for each object,and the robot motion path is simulated;Through TCP/IP Ethernet,the real-time and reliable communication mechanism between camera,PC software and industrial robot is studied and established.According to the extracted key feature points of workpiece edge,the machining traj ectory is generated,and the machining experiment of workpiece edge contour is completed.The developed platform meets the technical requirements in three aspects:matching success rate,trajectory accuracy and processing time.The system has the advantages of short software development cycle,high accuracy of image processing algorithm,fast communication mechanism and so on,which provides a reference for the wide application of hand eye cooperative working mode of industrial robots.