Research On Ingot Positioning System Based On Machine Vision

Machine vision is a technology that uses a camera to image the environment and uses a computer to automatically process images for useful information.The use of machine vision technology in industrial production can effectively reduce labor costs and improve the level of automation and intelligence.In this paper,combined with the needs of wafer production,a three-dimensional rotary measurement and positioning system based on line structure light is designed,which can effectively measure the geometric parameters of shaft parts.The main research contents of this paper are as follows:(1)Firstly,two schemes of visual measurement were compared: the triangulation method and the back projection method,and finally the triangulation method was chosen to construct the measurement system.Then the system is designed and the required hardware is selected according to the needs of the system.(2)The image preprocessing method was studied.It mainly includes image denoising algorithm and threshold segmentation algorithm to achieve the purpose of suppressing noise,removing background and obtaining strip area.After experimental comparison and research on various filtering methods,it is found that the median filtering can achieve the best results.The system finally uses median filtering as the image denoising method.Thresholding uses the largest inter-class variance method proposed by Otsu.This paper adds a coefficient to the Otsu method.The experimental results show that the effect is very satisfactory.(3)Research on line structure optical center line extraction algorithm,this paper studies several commonly used line structure light extraction algorithms,including extremum method,threshold method,gray center of gravity method,Hessian matrix method,Gaussian fitting,quadratic function The advantages and disadvantages of various methods are analyzed.Finally,the extraction precision of various methods is compared through experiments.The experimental results show that the Gaussian fitting method has better extraction precision and is suitable as the strip center extraction algorithm of this system.(4)In-depth study of camera pinhole model and calibration principle,and comparison of camera calibration procedures in MATLAB,Halcon,OpenCV,found that Halcon calibration accuracy is the highest,so this article chooses Halcon to calibrate the camera in the system.Then the calibration method of the laser plane is studied,and a method of fitting the laser plane using the calibration plate is proposed and implemented by Halcon and MATLAB.Finally,a rotating shaft calibration algorithm is proposed,which uses the least squares method and the calibration plate to realize the machining center of the machine.(5)Finally,the actual 3D measurement platform was built and tested.This article uses C++ language for programming,IDE uses VS2015,and uses the open source visual library OpenCV.In this paper,the cylindrical part is taken as the experimental object,and the particle swarm algorithm is used to complete the system construction.The results show that the need for wafer measurement can be met.