Research On TV Backplace Positioning Technology Based On Machine Vision

With the rapid development of economy and science and technology in China,the traditional manufacturing industry represented by household appliances is actively seeking transformation and upgrading to realize the intelligent production process.The production mode based on machine vision can improve production efficiency and reduce labor costs,and has been widely used in the production process of household appliances.In the process of TV production,it is necessary to spray glue on the TV back plate.Generally,the back plate is placed on the conveyor belt or two-dimensional plane,and the monocular vision is used to position the TV back plate to realize automatic glue spraying.This method is widely used because of its simple structure,low cost and easy implementation.However,with the improvement of life quality,large-size TV sets are more and more popular with consumers.The original single camera positioning method can not meet the existing production requirements of large-size TV backplane.Therefore,aiming at the problem of low positioning accuracy of monocular vision method in the glue spraying process of large-size TV backplane,this paper proposes a method of rapid positioning of plane large-size workpieces with dual cameras to meet the production requirements of existing large-size TV backplane.The main contents of this thesis are as follows:(1)The visual positioning system is selected.This paper analyzes and compares the two schemes of dual camera stereo positioning method and plane positioning method,and selects the plane positioning method as the positioning method of this system;According to the actual production requirements,the performance parameters of hardware equipment are discussed,the main hardware of the system is selected,and the production experiment platform of large-size TV backplane is successfully built.(2)Calibration of visual positioning system.The imaging principle of the camera is described and the imaging model is established.The camera is calibrated by Zhang Zhengyou calibration method,and the internal and external parameters of the camera are obtained.The calibration models of manipulator and camera are described,and the transformation relationship between image pixel coordinate system and robot coordinate system is obtained by using nine point calibration method.(3)In the visual positioning system image processing.Firstly,the original image is roughly trimmed,and then the region of interest(ROI)of the image is obtained by fine clipping through statistical method.Then the Gaussian filtering algorithm is used to preprocess the image to reduce noise interference.A horizontal and vertical edge detection method is proposed to extract the edge contour information of the TV backplane.Finally,the precise corner position of the extracted edge is obtained by the straight line fitting method based on Hough transform.(4)Realization and test of visual positioning system.This paper introduces the process of CAD reading path,analyzes the practice methods of single camera positioning and dual camera positioning,compares the similarities and differences between the two positioning methods,and obtains the transformation relationship between the TV backplane coordinate system and the robot coordinate system.On this basis,the two positioning methods are tested experimentally,and the system is analyzed from the three production indicators of rapidity,accuracy and stability,which verifies the feasibility and reliability of the dual camera positioning system.