Font Size: a A A

The Research And Application Of Vehicle Scene 3D Reconstruction Based On Binocular Stereo Vision

Posted on:2010-05-22Degree:MasterType:Thesis
Country:ChinaCandidate:Y HeFull Text:PDF
GTID:2178360275478049Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
The 3-D reconstruction technology of computer vision is a significant field of computer vision; it can help realizing 3-D untouched measurement and fast reconstruction. With in-depth study, the application of machine vision has spread to many fields, especially in Intelligent Transportation, it achieves speed development. Stake test is an important part of Driving License Examination; Study using autonomous mobile robot watch over the process of the examination, with computer vision technology, image process and pattern recognition. It has theory and application means to achieve Intelligent Stake Test.In this paper, the object of research is the vehicle scene in Automatic Stake Test System. Based on computer vision, research on using binocular stereo vision technology, to achieve 3-D measurement and reconstruction, in Automatic Stake Test background. To solve the problem that causes the wrong judge, owing to the loss of depth information. This paper does some jobs in Camera calibration, Stereo Match and 3-D reconstruction.Calibration: To get the precise camera parameters, the linear and non-linear methods were combined. First, Direct Linear Transform model was used to get the projection matrix. Then the matrix was decomposed based on the restraint condition, and the intrinsic parameters of camera were estimated. These estimates were fed to the non-linear optimizer. It's proved in the experimental that the computing speed of this algorithm is better than Zhang's algorithm; the accuracy is as good as Zhang's algorithm. Stereo Match: A stereo matching approach based on wavelet pyramid is proposed.Decompose the image to image pyramid with wavelet translation, let the matching point in low resolution as the restrict of matching in high resolution. To decrease incorrect matching and improve matching speed. The algorithm is verified on MATLAB, achieved dense depth map.3-D reconstruction: Achieve and process images with parallel binocular stereo vision system, to achieve the object ranges fast reconstruction with computation heavily reduced.The research indicate that Vehicle Scene 3-D Reconstruction based on Binocular Stereo Vision has features of rich information, direct viewing, that could improve the fair and objectivity of the Stake Test Examination.
Keywords/Search Tags:Computer vision, Camera calibration, Wavelet pyramid, Least square matching, Multi-resolution, 3-D Reconstruction
PDF Full Text Request
Related items