摘要： 提出一种基于简化无迹卡尔曼滤波(UKF)算法的摄像机标定方法。将平面靶标图像上的不同特征点坐标视为同一个特征点在不同时刻的运动坐标。为避免欧拉角描述法带来的奇异问题，用单位四元数描述世界坐标系和摄像机坐标系之间的变换关系，选取摄像机内外参数作为系统状态变量。结合实际应用背景，简化标准UKF算法，将其用于摄像机参数估计，在保证标定精度的前提下降低运算复杂度。仿真结果表明了该方法的有效性。

关键词: 摄像机标定, 平面靶标, 简化UKF算法

Abstract: This paper proposes a camera calibration method based on simplified Unscented Kalman Filtering(UKF) algorithm. The feature points in the planar target images are considered as motion coordinates of one feature point at different time. To avoid the singular problem from Euler angle description, the quaternion is used to represent the transform relation between world coordinate system and camera coordinate system. The intrinsic and extrinsic camera parameters are taken as system state variables. According to the application background, it simplifies the original UKF algorithm and uses it to estimate camera parameters, and reduces the calculation complexity while the calibration precision is kept. Simulation results show that this method is effective.

Key words: camera calibration, planar target, simplified Unscented Kalman Filtering(UKF) algorithm

中图分类号: 

• TP391