Design And Application Realization Of Robot Dynamic Assembly System Based On Vision

Vision-based industrial robot applications are a hot research topic in recent years for industrial automation production,and industrial robots with vision systems can effectively improve the intelligence level of robots.With the continuous improvement of computer technology and new research results of image processing technology,the traditional static vision application based on monocular camera is increasingly difficult to meet the current production needs,and the localization and grasping of dynamic targets has gradually become a hot spot for vision application in current industrial assembly line production.In this paper,the principle of recognition and grasping of dynamic targets by current industrial robots and the use of two monocular cameras to realize the visual recognition and positioning of static pallets and the dynamic following of dynamic targets on the conveyor belt are investigated by using the Chinese robot SCARA 6600.The conversion of the robot coordinate system to the static pallet and conveyor coordinate system,the robot's positioning and grasping of static targets,and the identification and following assembly of moving targets on the conveyor belt are studied,and a complete static grasping and dynamic following assembly system is constructed.The specific research contents and results are as follows.(1)Research and analysis of the current domestic and foreign research status of industrial robot vision applications,the current domestic and foreign research status of robot vision-based conveyor belt following and the current domestic and foreign research status of image processing software,and based on these status quo leads to the research content and difficult problems of this topic;design a dynamic vision assembly system based on the actual project requirements for the HASCO III SCARA robot,through analysis and The working principle of the system and the hardware equipment selection were studied,and the experimental test platform was built.(2)The model of monocular camera imaging system is studied and analyzed,the position conversion relationship between the base coordinate system of the robot,the coordinate system of the static material tray,the coordinate system of the conveyor belt and the coordinate system of the camera image is derived,and the matrix solution method based on the least squares method is studied on the basis of the traditional nine-point method hand-eye calibration algorithm;the operation flow and working principle of the following process package used in the subject are studied,and the robot The operation process and working principle of the following process package used in the project were studied,and the calibration of the workpiece coordinate system was realized,and the pulse equivalent calibration of the conveyor belt and the external encoder was also realized.(3)The image processing and stencil matching algorithm of the vision solution of the subject is studied and analyzed.The Hikvision Vision Master algorithm platform is used for image acquisition and recognition processing of the target object to be recognized,and on this basis,secondary development is done to realize the recognition and positioning of the heart-shaped ornament on the static material tray and the positioning of the greeting card to be assembled on the conveyor belt by using the absolute value encoder to obtain pulse feedback.By analyzing the relationship between the material to be recognized and the robot coordinate system,a reference value setting function based on the visual recognition feature point,the robot demonstration gripping point and the angle change of the visual image is designed to realize the offset based on the angle change of the recognition feature object,and the production process of the static material tray absorption and the conveyor belt dynamic with the assembly is designed.(4)QT4.8.5 + VS2010 platform was used to develop the vision application layer software of the subject,the robot interaction module was developed by using the secondary development interface of Huadong III,the image acquisition and processing module was developed by using the secondary development interface of Haikang Vision Master algorithm platform,the hand-eye calibration module was developed by using the principle of the nine-point method hand-eye calibration algorithm and analyzing the offset processing based on the angle change Through the experimental test platform,the vision application layer software was applied to the static pallet to absorb materials and the conveyor to dynamically follow the assembled materials to verify the correctness of the experimental principle,and through the analysis of the collected test data,a method based on the robot end-effector tool coordinate error compensation was designed.Finally,the whole static gripping and dynamic following assembly system was realized through experimental testing and it meets the project application requirements.