To detect the position of the camera from a sequence of un-calibrated images
Jaeseung Yu, Taehoon Kang, Kwanghee Choi, Mingi Kim · ITC-CSCC :International Technical Conference on Circuits Systems, Computers and Communications · 2007
This paper presents a method of detecting the camera position from an un-calibrated image sequence. In Epipolar geometry, the fundamental matrix provides some geometric information including translation and rotation of the cameras. But the information is inexplicitly encapsulated and does not have the scale factor which raises a problem about the displacement or the direction of the sequential motion of the camera. In this paper, the scale of the direction of each camera movement is synchronized comparing the camera matrix which is from the fundamental matrix of some two images with the others. Furthermore, scale factors are determined using the knowledge of the displacement which is made by the first motion of the camera. And to achieve good results the fundamental matrix is computed by the Levenberg-arquardt algorithm using the Direct Linear Transformation algorithm to find the initial value. Experimental results show that the track of the camera is accurately estimated.