Research And Implementation Of Three-dimensional Reconstruction Based On OpenCV

Vision is the important way to observe and recognize the world. People obtain andunderstand the vision information by our eyes and brain．Computer vision whichpercept and explain the3D information of the object world，tries to achieve the functionof vision like people’s vision by computer．The goal of computer vision is to make thecomputer have the ability to recognize the3D image，including the shape，position，posture，movement．The contents of computer vision is very widely， including stereovision，movement vision，color vision，initiative vision etc．Inferring3D configurationof objects from two or multiple images remains an important and difficult problem incomputer vision．The significance of3D reconstruction in many projects is not only intheory，but also in practice’s numbers of approaches for inferring3D configuration havebeen suggested，but computer vision is still in its initial stage．The difficulty arises，ofcourse，from the fact that two images are2D projections of the object and they have notthe adequate depth information of the object．As the development of manufacturingautomation，design automation，vehicle guide，intelligent robots and facial recognition，it is imperative to propose an efficient，exact，robust，general3D recover system．Stereovision is the important branch of computer vision，and the binocular vision is the hotspot of stereo vision．The fundamental principle of computer vision with dual．camerashas a strong resemblance to the visual perception procedure of mankind．We can getdifferent images with two different view points from two different cameras or twodifferent directions through the movements of one camera，then calculate the positiondeviation of images pixels based on triangular measure．Finally we can reconstruct the3．Dimensional coordinate of the object． Using this depth information we canreconstruct the surface of the object’s complete3D reconstruction system contents sixparts：image collection，camera calibration，image processing and feature extraction，image match，sequent processing．And this paper is focused on camera calibration，feature extraction and image match．This thesis mainly discusses stereo vision algorithmbased on OpenCV，Image acquisition is realized through CCD or digital camera and aimage acquisition card．which is a kind of passive methods．Generally speaking，a3Dreconstruction system consists of camera calibration，feature detecting，feature matchingand3D model calculation．The key issues is how to find out the good features andmatch them．in3D reconstruction are relationship between the different images．thispaper introduce the theory of binocular vision，then describe the camera calibration．To resolve the most difficult point of binocular vision--Sparse Points and DensePoints．The key points of this paper is camera calibration，image processing and imagematch．Design a three-dimension system and analysis the inaccuracy of the cameracalibration．The algorithm of this paper was realized by using OpenCV and its powerfulfunction．...