Research On Pearl Shape Sorting System Based On Machine Vision

With the continuous progress of machine vision technology and robot technology,robots gradually have the ability to reproduce part of the manual work.In the process of Pearl sorting,pearl shape classification mainly depends on manual work.However,each technician can only judge and sort one pearl at the same time.The fatigue work caused by a large number of sorting will lead to the reduction of sorting accuracy and efficiency.In order to solve this problem,a pearl shape sorting system based on machine vision is proposed in this thesis.The image of several pearls randomly placed is obtained by industrial camera.The shape of pearls is classified and located by computer,and the robot is guided to complete pearl sorting.The main research contents of this thesis are as follows.(1)Based on the analysis of Pearl contour characteristics,optical characteristics of pearl surface and Pearl sorting process,a pearl shape sorting system based on machine vision is designed.The hardware of industrial camera,lens,light source and lighting mode is selected.(2)In order to solve the problem of unclear contour and noise interference of Pearl image directly obtained by industrial camera,image enhancement technology is used to improve the contrast between Pearl edge contour and background,and image filtering algorithm is used to improve the image quality.This thesis analyzes the problem that some pearls touch each other,uses morphological method to process the filtered pearl image,optimizes the Pearl region,improves the traditional watershed method,realizes the fast and accurate segmentation of pearls with different contact degrees,and detects the edge of Pearl image.According to the shape characteristics of pearls,the model of pearl shape parameters is established,and the coordinates of Pearl centroid are calculated.(3)In this thesis,the sorting robot is simplified as a connecting rod mechanism,and the coordinate system and kinematics equation are established by D-H method.Matlab robot toolbox is used for kinematics analysis of forward and inverse motion,and Dobot robot model is established to realize the simulation and trajectory planning of sorting robot.The parameters of joint position,velocity and acceleration of sorting robot are analyzed,and the reliability of the model is verified.(4)In this thesis,based on the hardware selection,the experimental platform of Pearl sorting robot based on vision is built,and the hand eye calibration of vision system is completed.The Pearl with contact is identified,segmented,located,the shape parameters are calculated and the shape is graded,and the pearls with different shape grades are sorted.The experimental results show that the segmentation accuracy of this method is high,the shape classification effect is good based on multiple shape parameters,and the positioning and sorting accuracy of Pearl centroid is high,which meets the design requirements of the system.