Research On The Visual Positioning Technology Of Six Degree Of Freedom Industrial Robot

2020 is an unusual year.At the beginning of the New Year,a sudden epidemic made everyone experience the natural disasters and human disasters.The large-scale shutdown of the enterprise caused everyone to stay at home,and the development of the manufacturing industry once declined,causing a decline in the overall social economy.The country's strong epidemic prevention and control has made this epidemic prevention and control battle about to win.At present,enterprises are gradually resuming production,so at the same time,the personal protection of workers is particularly important.Compared with humans,industrial robots will not be infected with terrible viruses in the face of the epidemic,and enterprises will not stop production and production on a large scale,and will not have such a huge impact on society.However,traditional teaching-type industrial robots require workers to accurately teach one by one according to the position of the workpiece in advance,which has strong limitations.With the development of artificial intelligence technology in recent years,the fusion of machine vision and industrial robot technology can enable the robot to identify,detect,and locate the target workpiece,and can perceive the position,motion and other information of its working environment,greatly improving the industry The degree of robot intelligence has become the current research focus of industrial robots.This subject is aimed at the KUKA KR10-R1420 industrial robot as the research object on the industrial 4.0 intelligent manufacturing production line training base.The machine vision and industrial robot kinematics control are combined.Grasp the task of the target workpiece,and then carry out the research on the key technology of six-degree-of-freedom industrial robot visual positioning.To complete this project,the main research steps of this article are as follows:First,a brief introduction to the control system configuration of asix-degree-of-freedom industrial robot vision positioning system,a brief description of the mechanical structure of the KUKA KR10-R1420 industrial robot and the establishment of a link coordinate system and kinematics equations on the improved DH theory The positive solution and the inverse solution,and a model is established in Matlab to verify the kinematics forward solution and inverse solution.Then,in-depth research on camera calibration technology and Eye-in-Hand hand-eye calibration technology,analyze the basic principles of the most commonly used small hole imaging camera,deduces the conversion of pixel coordinates of the object in the image to the actual coordinates in three-dimensional space The coordinate conversion from the pixel plane to the image plane,the coordinate conversion between the image plane and the camera coordinate system,the coordinate conversion between the camera coordinate system and the world coordinate system,and finally the coordinate conversion between the pixel plane and the world coordinate system.Based on Zhang Zhengyou's method and Matlab camera calibration toolbox,calibration experiments are performed,and the camera's internal and external parameters are calculated to complete the camera calibration.Analyze the calibration method and ideas of Eye-in-Hand hand-eye system,and derive the process of solving AX = XB of Eye-in-Hand configuration.Finally,the image collected by the camera is preprocessed,and the methods of image transformation,image noise reduction,and image enhancement are analyzed through experiments.The method principles of mean filtering,Gaussian filtering,median filtering,and adaptive filtering are studied in detail.The median filtering algorithm is selected as the image noise reduction processing method in this study.A comparative experiment was carried out on five image edge feature detection algorithms,the advantages and disadvantages of each algorithm were analyzed,Canny algorithm was selected as the image edge feature detection algorithm in this study,and finally the image processing scheme of the monocular vision positioning system was determined.The contour of the nut is approximated to regular hexagon by image processing,and the edge feature matchingalgorithm based on Hu moment is used to identify the nut.In order to represent the similarity between the recognized workpiece and the target template,the concept of similarity is introduced,and then the hexagonal geometric center is used to determine the position of the nut in the world coordinate system,and finally the industrial robot is controlled to grasp the target workpiece.