数码相机标定方法研究

Research on Digital Camera Calibration Method

数码相机在计算机视觉中的应用逐渐普及和深入，在应用中数码相机的标定是相当重要的。根据数码相机的特点，提出了一种新的标定方法。针对数码相机的某一状态，先确定其内部参数矩阵，再通过采集的图像求出外部参数矩阵。首先建立了相机成像的几何模型，并将此模型分解成内、外参数矩阵。文中详细介绍了数码相机内部参数矩阵中的各元素所对应的物理参数以及求解它们的原理和方法，从而建立内部参数矩阵。推导了求解外部参数矩阵的计算公式，并且介绍了相应的求解方法，再由外部参数矩阵求解出对应的物理参数．通过实验表明这种方法有相当高的精度，且操作方便。

Digital camera is being widely used in computer vision. Digital camera calibration is a preliminary step toward computational vision. This paper brings forward a new method of the digital camera calibration in terms of its feature. First, intrinsic camera parameter matrix is solved under a state. Then, extrinsic camera parameter matrix of images is solved. Geometric model of digital camera is built and decomposed to intrinsic and extrinsic parameter matrixes. In the paper, each element of the intrinsic parameter matrix and the theory, and method for solving them are introduced. The intrinsic parameter matrix can be got. The formula for solving extrinsic parameter matrix is derived and the method for solving them is introduced. Extrinsic physical parameters can be solved by extrinsic parameter matrix. Experiment proves that the method is accurate and convenient.