Research And Design Of Intelligent Sorting System Of Production Line Based On Machine Vision

With the rapid development of industrial technology,the requirements for the intelligence and automation of manufacturing have become higher and higher.Machine vision technology has become a powerful weapon for the country to realize Industry 4.0,and has been widely used in packaging,logistics and other industries.At the same time,workpiece sorting is an important part of intelligent manufacturing.The purpose is different from the task requirements of large batches of scattered workpieces on the sorting assembly line.The traditional sorting method cannot meet the requirements of intelligence.Therefore,the introduction of machine vision technology is important for solving the problem of automatic workpieces.The sorting problem is of great significance to improve the efficiency of industrial production.This paper designs and develops machine vision technology recognition and intelligent sorting system,and conducts experimental research.The main research contents and results of this article are as follows:(1)Complete the structural design of the machine vision sorting system.Design a machine vision system based on 3D camera and 2D camera,and combine motion control module to sort scattered workpieces.Through inquiring data and calculations,the parameters of the hardware equipment are determined,and the type selection is performed.The interface design of the hardware equipment is completed to ensure the efficient and smooth operation of the system.(2)System calibration is carried out for the entire machine vision system,which mainly covers camera calibration and hand-eye calibration.Use the HALCON machine vision software to calibrate the camera of the system,and then obtain the internal and external parameters of the camera.The eye-in-hand connection method is adopted for the hand-eye calibration between the camera and the manipulator,and the camera coordinate system is converted to the end tool coordinate system of the manipulator.,And then realize the conversion between image coordinates and world coordinates.(3)The pose estimation algorithm of the scattered workpiece is designed.First,the depth map collected by the 3D camera is converted into a point cloud model,and then the point cloud information is mapped into X image,Y image and Z image,and the Z image is calculated by the region growing method.Perform image segmentation,extract the region of interest of the highest position workpiece,and finally use the region of interest of the 2D image to extract the region of interest in the 3D point cloud model,and convert the obtained workpiece pose into the space coordinates of the robot base coordinate system.Complete the robot’s gripping task.(4)A workpiece recognition and positioning algorithm based on a 2D camera is designed.First,smoothing is performed,and the different processing effects of the three filters are compared through experiments.The median filter is determined as the main method of denoising.Second,the region of interest is extracted by Blob analysis,and then Obtain the edge information of the target workpiece,and make a template according to the edge information of the target workpiece,use the shape-based template matching algorithm to determine the type of the workpiece,and complete the sorting.(5)Design algorithm compilation and software development based on sorting system,and realize sorting of workpieces.First,the overall process of the sorting system is planned and explained,which is mainly divided into two parts: the grabbing of scattered workpieces and the target recognition and positioning,and they are described in detail.Then,based on the joint programming of MFC and HALCON in C++,the software for grasping scattered workpieces and the software for target recognition and positioning were developed.Finally,the experiment was verified.