Camera Calibration Algorithms Using 2D and 3D Calibration Field

As well as the digital camera comes about, 3D reconstruction becomes simple and economical. Using general camera's photo to reconstruct 3D scene, the first step is cameral calibration. In research field of computer vision, direct linear transformation (DLT) is a common algorithm for camera calibration, but the calibration accuracy of DLT is limited. In the field of photogrammetry, self calibration is a high precision method for camera calibration, but this method needs initial value of camera parameters. The current research focus in camera calibration is binding DLT and self calibration, DLT algorithm calculates the initial values and least square self calibration calculate the accuracy parameters. The result of DLT is a homograph matrix, the problem of binding DLT and self calibration is how to get initial camera parameters from DLT result. Paper main introduced the formulas from DLT result based 2D and 3D calibration field to initial values of camera parameters and simply discussed the camera calibration method.