Research And Application Of Industrial Robot For Sorting Based On Image Processing Technology

In the past decade,artificial laborers working on production lines to sort components has increasingly been replaced by industrial robots in the field of manufacturing and processing.In factories today,most of the industrial robots were used working based on teaching and off-line programming.If the position,size,and environment of parts change during the sorting process,the robots need to be taught or programmed again.To solve this problem,an image processing technology is applied to industrial robots to make the working process faster and more intelligent.In this paper,the Evert ER3A-C60 industrial robots and industrial cameras were used to build the experimental platform,and an industrial robot for sorting system based on image processing was established.First,the industrial robots,industrial cameras and the objects on experimental platforms were identified as research objects.The global mathematical model between the world coordinate system,camera coordinate system and image coordinate system were constructed and the transformation matrix of the coordinate system was also calculated.Then,the image of the object on the platform was obtained through the camera,and the graying and binarization processing methods was used to process the image.Gaussian low-pass filtering,median filtering,and bilateral filtering were used separately on the processed images to obtain clearer image.On the basis of preprocessing,the combination of template matching algorithm,Canny edge feature extraction and contour moment were used to complete the recognition and positioning of parts.After that,the position coordinates of the target object were obtained and sent to the robot.Finally,the D-H parameters of the Evert ER3A-C60 industrial robot were calculated to analyze the kinematics and inverse kinematics.The joint trajectories of the robot were obtained using the robot toolbox in Matlab and the UG's motion simulation.By comparing the advantages of the two trajectory planning methods,the motion of the robot was guaranteed to be smoother and more stable.After analyzing the corresponding theoretical basis,the sorting experiment was designed and completed on the experimental platform.During the experiment,there was a relatively small deviation between the absolute position of the robot and the position of the image acquisition,which reason for the analysis was due to external factors such as the absolute deviation of the robot's installation,the verticality deviation from the workbench during camera installation and so on.Based on the above reasons,error statistics were used to compensate.Through multiple experiments,each deviation was recorded to calculate the average error value,and the resulting error was directly compensated to the upper computer control program.The experimental results showed that the success rate of sorting and picking of parts were above ninety percent.