Research And Design Of Workpiece Sorting System Based On Machine Vision

With the continuous advancement of intelligent manufacturing technology,robots are playing an increasingly important role in manufacturing production.Sorting robots are a type of industrial robots,which can replace workers to sort workpieces on the assembly line.The traditional sorting robot plan the grasping direction and movement path by teaching or offline programming,although it can complete the sorting operation,this type of sorting robot has strict requirements on the placement position and size of the workpiece.If the placement position and size of the workpiece to be sorted change,it is easy to cause the robot to fail to grasp the workpiece.Sorting robots based on machine vision can detect and sort workpieces in real time,and have the advantages of high flexibility and stable sorting.According to the demand of domestic manufacturing market,the research on the workpiece sorting system base on machine vision has certain application value and practical significance.This dissertation proposed a set of vision positioning and sorting schemes for the sorting requirements of multiple workpieces on the conveyor belt,and builted an experimental platform composed of four parts: conveyor belt unit、image acquisition unit、image processing unit and robot sorting unit,and realized the function of identification,positioning and sorting of various of workpieces.The main work and research contents of this dissertation included the following aspects:(1)Designed and realized a vision-based workpiece sorting system.The actual requirements of the sorting system were analyzed,and the hardwares were selected according to the function and actual environment,and the vision-based workpiece sorting system was designed.Industrial cameras were used to obtain the workpiece image on the conveyor belt in real time.Through the vision technology,the position of the workpiece was detected and located,and then the Cartesian coordinate robot was controlled by the computer to realize the grasping and placement,so as to realize the real-time multi-workpiece sorting.(2)Studied the calibration technology of the visual sorting system.The related theories of the camera imaging model were analyzed,and the camera calibration experiments were carried out based on Matlab and Halcon respectively.According to the calibration results,Halcon was selected for system camera calibration.Due to the fact that the actual imaging model of the system was a distorted image model due to lens distortion,the polynomial distortion model based on Halcon was adopted to correct the radial distortion and tangential distortion.Through the camera calibration and hand-eye calibration process,the conversion relationship between the position of the target object in the image and the position of the target object in the objective world was determined.(3)Researched and realized the target positioning in the visual sorting system.The image preprocessing technology was introduced,and the image of the workpiece was preprocessed.The image segmentation method based on threshold was used to binarize the workpiece image,and the image morphology method was used to eliminate the noise and holes in the binary image.The template matching method and theory were studied,and the advantages and disadvantages of different matching methods were compared,according to the actual contour features of the workpiece,the template matching method based on shape and the hierarchical search strategy based on image pyramid were adopted to match and locate the workpiece image.Combined with the transformation relationship between camera coordinate system and robot coordinate system,the relative coordinate information of workpiece in robot coordinate system was obtained.(4)Aiming at the complexity and time-consuming of the traditional motion positioning method,a moving workpiece positioning method oriented to the speed of conveyor belt was designed.The offset of the workpiece on the conveyor belt was calculated by the speed of the conveyor belt obtained by the wheel encoder,and the coordinate information obtained by the static workpiece positioning was added to obtain the position coordinates of the robot to be grasped.The sorting system designed in this dissertation used machine vision technology to identify and sort various kinds of workpieces on the conveyor belt,repeated experiments showed that the workpiece sorting system had the advantages of strong real-time and high sorting accuracy,and can be applied to the workpiece sorting task of industrial production line.