摘要
提出了一种基于Hough变换的航空摄像机外参数估计新方法。该方法使用共线条件方程描述摄像机成像关系,用图像与物方特征点之间的组合对外参数空间离散点进行投票,将得票最多的离散参数点确定为参数估计值。为解决共线条件方程的非线性以及外参数之间的相关性问题,采用迭代方式进行参数估计。与基于最小二乘原理的传统外参数估计方法不同,该方法不需要事先对图像与物方空间特征点进行配对,解决了传统最小二乘法的特征匹配难题,避免了最小二乘法对错误匹配的鲁棒性较差的缺点。分别使用仿真和真实航拍图像对算法的有效性进行了检验,结果表明:摄像机坐标估计误差小于15 m,姿态角误差小于1°。参数估计结果能够满足舰炮作战中对侦察摄像机外参数估计精度的要求。
A new approach for exterior camera parameters estimation based on Hough transform is presented. By using the collinear equations to model the camera and combining the feature points on the image plane and in the object space to set votes for a certain point in the discrete parameter space, the point with the maximum number of hit is defined the right parameter. An iterative estimation pro cedure is adopted to process the nonlinearity in the collinearity equations and the correlation of the ex terior parameters. Unlike the conventional methods, the new technique can estimate the exterior pa rameters without making object space. Two experi a correspondence between the feature points on the image plane and in the ments are carried out to examine the validity of the proposed algorithm. Ex perimental results show that the errors of the camera coordinates and the pose angles are less than 15 m and 1° respectively. The estimated parameter results can meet the requirements of the combat using ship borne guns.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2007年第8期1269-1274,共6页
Optics and Precision Engineering
基金
国防科技预研基金资助项目(No.40108060104)
