基于畸变分离的摄像机标定方法

在摄像机标定过程中,为了避免对摄像机模型中的畸变系数进行多次重复标定,提出一种将二阶径向畸变系数与摄像机模型分离的标定方法.该方法利用畸变形成的围线面积作为畸变评测函数,用模拟退火原理改进粒子群算法的惯性权重和学习因子;然后用改进的粒子群算法标定摄像机的畸变系数和图像中心点坐标,最后计算其他的摄像机参数.该方法无需预先知道摄像机的任何内外参数,算法简单,易于实现.实验表明,该方法与传统的非线性优化方法相比,图像坐标的平均反投影误差明显减小,而且具有更好的鲁棒性和精度.

Camera Calibration Method Based on Distortion Separation

In the process of camera calibration， in order to avoid repeating calibration of the distortion coefficient in camera model， a distortion separated camera calibration method was proposed. The second order radial distortion was considered， in which the area of contour line formed by distortion is utilized as the criterion， and the inertia weight and learning factor of the particle swarm optimization algorithm based simulated annealing were improved. Then the improved particle swarm optimization algorithm was utilized to calibrate the distortion coefficient and principal point coordinate of the camera. Finally， the other camera parameters were calculated. The proposed method was simple and easy to implement without needing any internal and external parameters of the camera in advance. Experiments show that the proposed method has lower mean back-projection error and better robustness compared with the traditional nonlinear optimization methods.