Performance evaluation of linear camera calibration techniques

Camera calibration represents an important step in 3U reconstruction process. A lot of linear camera calibration techniques exist which try to compute the camera intrinsic and extrinsic parameters with a close form solution. This close form solution arise form the minimization of a linear cost function. Since the noise in the points' coordinates, the computed parameters are an approximation of the real ones. Therefore, this cost function can include some restrictions in order to improve the robustness of the algorithm. In this paper a summary of the most important linear camera calibration techniques is done. Afterwards, an evaluation of its performance is carried out. This evaluation is done from the point of view of the parameters which have more influence in the results. These are the number of points and the noise level.