Research Of Visual Positioning System For Round Steel End Surfaces Of A Bundle

Round steel is the main product of the iron and steel industry,and it is also an important raw material for production and processing in the industrial field.For different industries,the quality of round steel requirements is also different.A wide variety of round steel is produced by the steel industry.In order to identify the basic information of round steel,labels are affixed on the end surfaces of round steel.The labels are now affixed to the end surfaces of round steel by hand.In order to improve efficiency and reduce labors intensity,industrial robots instead of manual labeling become the trend of development in the future.The key problem of robot labeling is to get the center coordinate of end surfaces of round steel.The key to solve this problem is to recognize and locate the end surfaces of round steel.So,it is a strong practical significance to research of visual positioning system for round steel end surfaces of a bundle.In this thesis,the possibility of visual positioning of round steel end surfaces was analyzed from two aspects of visual positioning method and image processing.Finally,the method of visual positioning and the process of image processing were determined.In the determination of the visual positioning scheme,monocular vision positioning method and binocular vision positioning method were analyzed.The main and auxiliary eye image recognition and positioning method was proposed,and the system was established.The main and auxiliary eye system contained a master camera and a secondary camera.The main camera was used to determine the X and Y coordinates of the round steel end surfaces.The stereo vision system composed of main camera and auxiliary camera was used to calculate the depth Z value of the round steel end surfaces.In the aspect of image processing,Hough transform and least square method fitting circle were compared.In order to ensure accuracy of the round steel end surfaces,the least square fitting circle was selected.The matching points were determined by centroid matching when the center coordinates were known.This method simplified the process of stereo matching.By applying the MATLAB software,positioning program of round steel end surfaces was written.The program included two parts: the camera calibration subroutine and the image processing subroutine,and realized the positioning of round steel end surfaces center.This thesis presented a vision positioning system for main eye and auxiliary eye,and determined the process of image processing.In order to ensure the accuracy,the X and Y values were calculated by the main camera,and the Z value was calculated by the main camera and the auxiliary camera.This made up for the disadvantage that Z values couldn't be captured by a single camera.In the process of image processing,the method of Z value calculation was simplified,so the efficiency was improved.In this thesis,the center coordinate of the end surfaces of round steel was divided into two parts.In this way,not only the accuracy was guaranteed,but also the calculation of depth was simplified.