Calibration Of Central Catadiontric Camera Using Projection Properties Of Lines

With the rapid development of science and technology,the technology of camera calibration should be continuously improved so as to keep up with the trend of social development.After analyzing,this thesis proposes an algorithm to calibrate central catadioptric camera by using space straight line.The idea of the algorithm is as follows: By the unit sphere projection model of central catadioptric cameras,when there are at least two unparallel of three straight line in the space,first the projection on the unit as the unit sphere can form three groups of antipodal points and corresponding on the image plane can form three groups of antipodal image points,according to the line of antipodal points will pass through the center of the unit visual sphere,and the line of the antipodal image points must pass through the principal point,so the principal point can be obtained;secondly,two methods to obtain vanishing lines are proposed,which are based on the properties of pole-polar and harmonic conjugate;then the image of the circular points is obtained by intersecting the conic and vanishing line;after the image of the circular points is obtained,the image of the absolute conic can be estimated by the least square method;finally,Cholesky was used to decompose the image of absolute conic to obtain the internal parameters of the camera.When all three lines in space are parallel and the lines are perpendicular to the optical axis of the camera,the algorithm in this paper can also be used for calibration.The modified absolute conic can be obtained from the common pole-polar of the image of lines,and the mirror parameters can be obtained further.The calibrator used in this thesis is a checkerboard.After analysis,we can estimate the external parameters and distortion coefficient of the camera according to the corresponding points.In order to improve the accuracy of the experimental results,this thesis also optimizes the image of the straight line,because the algorithm in the full text is based on the line image which has been fitted.A large number of simulation experiments and real experiments are carried out in this thesis,and the 3D reconstruction experiment is also carried out in the real experiment.The results show that the algorithm is feasible.