基于位姿约束的大视场双目视觉标定算法

发布时间：2018-07-17 03:27

【摘要】：为了提高大视场、远距离的双目摄像机标定精度,提出一种基于位姿约束的摄像机标定算法。该方法利用双目摄像机之间的三维位姿关系是刚体变换这一属性,标定出左、右摄像机相对位姿的外部参数。利用相对位姿为约束条件求取摄像机的初始内部参数,剔除较大的重投影误差值对应的标定图像组,重复迭代直至重投影误差平均值小于指定值,得到多个待优化的摄像机内部参数。再将最后标定图像组的角点坐标、待优化的摄像机内部参数和相应的外部参数,建立一个以角点三维重构坐标值与实际设定角点三维坐标值的模均值为最小的目标函数,求解出双目摄像机标定参数的最优解。该方法很好地解决了误差大的标定图像造成的影响,且充分利用了双目摄像机之间的位姿约束关系。通过仿真和标定实验可以看出,本文方法可以实现大视场双目摄像机的高精度标定。
[Abstract]:A camera calibration algorithm based on pose constraints is proposed to improve the calibration accuracy of binocular cameras with large field of view and long distance. In this method, the external parameters of the relative pose of the left and right cameras are calibrated using the attribute that the three-dimensional pose relationship between the binocular cameras is rigid body transformation. The initial internal parameters of the camera are obtained by using the relative pose as the constraint condition, and the calibration image groups corresponding to the larger reprojection error are eliminated, and the iterative process is repeated until the average value of the reprojection error is less than the specified value. Multiple camera parameters to be optimized are obtained. Finally, the corner coordinates of the image group are calibrated, the camera internal parameters and the corresponding external parameters are optimized, and an objective function is established, in which the mean value of the reconstruction coordinates of the corner points and the real values of the three-dimensional coordinates of the corner points are the minimum. The optimal solution of calibration parameters of binocular camera is obtained. The method solves the influence of the calibration image with large error and makes full use of the pose constraint relationship between binocular cameras. Through simulation and calibration experiments, it can be seen that this method can achieve high precision calibration of large field binocular camera.
【作者单位】：长春理工大学计算机科学技术学院;
【基金】：国家科技支撑计划重大项目课题(2012BAF12B22) 吉林省重大科技攻关项目(2012ZDGG004) 吉林省科技发展计划项目(20130303011GX);吉林省科技发展计划项目(20140204050GX) 长春市科技计划项目(14KG013)
【分类号】：TP391.41

【相似文献】