A new method for improving camera calibration precision

The camera calibration precision directly affects the computer vision application. A method for improving camera calibration precision is proposed in this paper, which follows in details: Firstly, a new iterative sub-pixel detection approach based on the Forstner operator is proposed. Secondly, instead of print plane, this paper uses LCD (completely flat Liquid Crystal Display) for camera calibration. Finally, a new distortion correction method using image corners is proposed, which is based on the fact that the straight line is still straight in projective transformation. This method could be used to compute the radial distortion coefficients, and correct the corners location. Both simulation and real experiments show the effectiveness of the proposed method.