Research On Calibration Of Camera Array

With a larger field of view, camera array is widely used in virtual reality, surveillance, image-based rendering and 3D reconstruction. The camera calibration is an indispensable step for recovering 3D information from 2D images and provides the underlying basis for object reconstruction and recognition. The current study on single camera calibration is relatively mature, and can get good results. Therefore, calibration of camera array based on the single camera calibration method has important theoretical significance and practical value.In order to calibrate camera arrays, we choose appropriate single camera calibration method, and extend it to the camera array calibration. Each camera is calibrated based on planar-based calibration method, and then, the relative positions between all the cameras are calibrated.Here are the details of several parts mentioned above:1) Through studying the main methods of camera calibration, we choose planar-based calibration method to calibrate single camera, and the results are good.2) Using the homography matrix between planar and image plane, all the camera's projection matrix and the template parameters are obtained via factorization. Then, the results of factorization are transformed into the same reference coordinate system to get relative position between cameras.3) We propose a stepwise refinement camera array calibration method. Select a camera as a reference, and then the relative position of other cameras to the reference camera are obtained step by step. Finally, we can get all cameras'position in the world coordinate system .The calibration results show that the proposed method can obtain accurate internal and external camera parameters and relative position between cameras.