Camera Calibration And 3d Reconstruction Techniques To Study

3D reconstruction is one of the basic tasks of computer vision.It computes geometric information of the object in the 3D space from the 2D image taked by camera, then reconstruct and recognize the object.The relation,between the 3D geometric positions of the points on the surface of object and the corresponding point of the 2D image,is decided by camera's imaging model,and these geometric model parameters is called camera parameters.The process of computing these parameters is called camera calibration.The precision of the calibration results directly affects the accuracy of 3D reconstruction.An algorithm of camera calibration based on circular markers of different size is put forward.Firstly the calibration plane is designed which have the circular markers.Then the method of detecting the markers and corresponding them is presented.Finally computing the camera parameters based on zhang's planar calibration algorithm.The experimetal results show that the method is correct.This thesis studies the recognition and matching of the characteristics of image.An method of reconstructing the poses of cameras and reversly get the three-dimensional structure is put forward.First,calculating the basic matrix and fundamental matrix based on the camera parameters we have got.Then the poses of cameras is computed.Finally, the coordinates of three-dimensional points is calculated.The experimetal results show the correctness of the method.