Font Size: a A A

Research And Application Of 3D Digitization Technology Based On Binocular Vision

Posted on:2018-06-21Degree:MasterType:Thesis
Country:ChinaCandidate:C LiFull Text:PDF
GTID:2428330515455868Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
Binocular stereo vision is a popular direction in the field of computer vision research nowadays.It simulates the visual function of human eyes through the binocular cameras.The computer image information is processed by computer to realize the recognition and understanding of the objective world space scene.Today it is widely used in many areas of modern society,such as vision navigation of robot,aerial survey of unmanned aerial vehicles,medical surgical navigation,three-dimensional non-contact measurement,virtual reality and so on.Therefore,the research and application of 3d digitization of spatial objects based on binocular stereo vision has important theoretical and practical value.This paper focuses on three major aspects of camera calibration technology,image matching technology and 3D coordinate reduction technology in binocular stereoscopic 3D digitalization.The main work of this paper includes the following parts:(1)Camera calibration.In this paper,Zhang Zhengyou's calibration method is used to carry out the calibration work of the camera,and the calibration of the binocular camera and the calibration of the binocular camera based on the Matlab calibration toolbox and OpenCV are carried out.The image is corrected according to the calibration results.The accuracy of the calibration procedure in terms of calibration efficiency and re-projection error.(2)Stereo matching.In this paper,the matching operation of image pair under SIFT and SURF is realized,and the random matching of the initial matching relation is effectively eliminated by the random sampling consistency algorithm(RANSAC).The two methods of SIFT and SURF are evaluated by mis-matching Characteristics of the effectiveness of three-dimensional matching.(3)Three-dimensional digital.Based on the parallel optical axis model of the traditional binocular camera and the depth information of the intersecting optical axis model,this paper proposes an algorithm to recover the depth information of the space object in any position of binocular camera,and completes the three-dimensional digitization of the target object Work,combined with triangulation,texture mapping and other reconstruction of the target object of the three-dimensional digital model.From the final experimental results,the stereoscopic constructed platform accurately extracts the characteristic information of the surface of the object,completes the matching between the image pairs,and restores the depth information of the target object,combined with the triangulation of the subsequent three-dimensional reconstruction process,texture mapping and other visual work,better reconstruction of the target object of the spatial structure and geometric shape.
Keywords/Search Tags:Camera calibration, Stereo matching, Depth information extraction, Three-dimensional digitization
PDF Full Text Request
Related items