Calibration Of Central Catadioptric Cameras Using Veronese Mapping By The Space Checkerboards

The camera calibration is an essential step in computer vision, whereas calibration of central catadioptric camera is relatively mature, the result is more accurate, and it also has a wide range of applications in three-dimensional reconstruction, three-dimensional measurement, robot navigation,visual surveillance and so on. The more accurate results of calibration, the more specific and rich information of the three-dimensional space we get, which is more beneficial for the application of computer vision.The appropriate template and calibration method have a great impact on the calibration result, and in this thesis,we proposed a camera calibration method using the space checkerboards as a template that having 3 planar faces which are perpendicular to each other, using the linear method to obtain the intrinsic parameters of camera, and to estimate the extrinsic parameters. The method template is simple and easy operation,and it is effective for every central catadioptric system. The main contents are as follows:1. In the central catadioptric camera, using Veronese mapping to solve projection matrix between the checkerboard and the image plane;2. Analyzing of the relationship between the camera projection matrix with intrinsic parameters of camera and the extrinsic parameters. Solve camera intrinsic parameters by a linear method; Estimating the extrinsic parameters of camera from the projection matrix,and followed by a non-linear optimization to it. lasely,we obtain the rotation matrix by reduct dimension.3. Using matlab to simulation and real experiment, verify the feasibility and stability of the algorithm; and as a whole to analysed the experimental data;4. Using the intrinsic parameters to correct image and to restore the perspective image; Detecting the accuracy by Hough transform.