Non-metric Digital Camera Calibration Based On Vanish Point Geometry And Radialalignment Constrain

Camera calibration, the key technology and top priority for three-dimensional information extraction from two-dimensional digital images, is widely employed in photogrammetry measurement, machine vision, security monitoring, three-dimensional reconstruction, biomedicine application and remote sensing mapping as well as na-vigation control and so forth. In this paper, one non-metric digital camera single calibration method with high precision was developed and tested while taking two-di-mensional square grid as the reference. The main tasks are as follows:Firstly, the existing digital camera calibration techniques were concluded and sum- marized, in addition, three calibration algorithms based on two-dimensional reference plane, such as TSAI two-step calibration, two-dimensional DLT calibration and camera calibration with vanishing points were highlighted and validated. At last, the advantages and disadvantages of these three algorithms were analyzed and verified through experiments.Secondly, learning the basic principle of TSAI two-step calibration and camera calibration with vanishing points and using planar grid as a frame of reference, the strict mathematical relationships between dual vanishing points formed by perspective projection of two orthogonal straight linear and parameters in the process of TSAI two-step camera calibration were deduced and established based on Vanish Point (VP) geometry and Radial Alignment Constrain (RAC). Then the nonlinear calibration equation was established, which was solved by LS criterion for optimal solution in order to achieve the precese solution of the interior and exterior orientation parameters of camera and radial distortion parameters.Finally, the improved algorithm proposed in this paper was compared with the existing algorithms using the ideal simulation images and real images to verify the feasibility and high accuracy of the improved algorithm.