Video sequence calibration and applications

Abstract

In this paper we propose a new technique to calibrate a video sequence. We will compute the location and rotation of the camera for each frame of the video sequence. Classical methods for camera calibration work properly in the case of a small number of cameras. When we take an image sequence with a video rate, the number of frames is very large and classical methods do not work properly. To avoid this drawback, we focus the calibration process in the estimation of the coordinates of a set of 3D points in the scene, so the unknown are these 3D point coordinates rather than the projection matrix associated to each frame. The advantage of this approach is that with a few number of 3D point coordinates we can estimate the projection matrix of any frame using the relation between the 3D points and its projection on each frame. In order to compute such set of 3D points, we will use a classical camera calibration technique applied on a small subsequence of frames taken from the large original video sequence. We also perform an iterative procedure to include the information of all the cameras in the calibration computation. To illustrate the capabilities of the proposed method, we will apply this technique to include virtual 3D objects in a real video sequence.