Research On Three-Dimensional Measurement And Visualization Technology Based On Binocular Vision

Machine vision is a new multidisciplinary research direction,with the help of rapid progress in computer vision and image processing technology,whether in theoretical research or applied fields,it has achieved unprecedented development.Binocular stereo vision is an important research hotspot of machine vision.It has the advantages of high automation,easy system construction,real-time online detection.Today,binocular vision technology has been widely used in many fields such as industrial inspection,robot vision navigation,etc.The research of this technology is of great significance.This article takes a flange workpiece as the research object,using the binocular vision technology to complete the three-dimensional measurement of the workpiece and the three-dimensional model visualization.The specific research content is as follows:In terms of camera calibration,this paper studies the Zhang's calibration algorithm.When binocular measurement is performed,the calibration of the camera must first be completed.Therefore,the study of the calibration method is very critical.This paper uses Zhang's calibration method to calculate the camera's internal and external parameters and distortion coefficients in a binocular measurement system to obtain higher accuracy.In the aspect of image preprocessing,the denoising method and correction method are studied in this paper.Workpiece images often introduce noise during the acquisition process.Through different airspace filtering methods,the images are preprocessed to determine that the Gaussian filtering algorithm can effectively remove the noise in the image.Before feature point matching,in order to make both left and right image feature points on the same image line,this paper adopts epipolar line rectification method to rectify the workpiece image.In feature matching,when matching feature points based on SURF feature matching algorithm,there will be mis-match problems.In this paper,an improved SURF featurebased matching algorithm is designed to processe the images in the standard image library and the binocular system respectively.Experimental results show that the improved algorithm improves the matching efficiency of feature points.In 3D measurement,in order to complete the measurement of the 3D geometric information,and to display the three-dimensional structure of the workpiece visually,this paper uses the least squares method to solve the three-dimensional point coordinate value,so as to obtain the data of the workpiece and visualize it through MATLAB's excellent graph function.In terms of system testing,this paper designed the software interface of the binocular measurement system,integrated the 3D measurement visualization technology,and tested it with the laboratory binocular hardware platform.Through analyzing the measurement of the size of the flange workpiece,it was proved that the Measure the effectiveness of the system.