利用单幅影像的空间目标姿态测定方法

Pose Determination of Space Object with Single Image

提出一种利用已知外方位元素的单幅影像测定空间目标姿态的方法.在观测到目标物体上一定数量控制点的情况下,利用角锥体方法求解本体坐标系下影像外方位元素的近似值,通过坐标变换获得物方坐标系下的目标姿态参数初始值并带入共线方程,采用最小二乘平方差解算出精确姿态参数.该方法避免了利用2台或多台摄像机进行运动目标姿态测定的摄影同步问题,具有更大的灵活性和实用性.模拟和实际数据的试验结果验证了方法的正确性.试验结果还表明,摄影中心距离目标越远,相机的焦距越长,姿态测定的精度就越差.实际应用中可以根据具体情况和精度要求,灵活选择不同数量的控制点和合适焦距的相机来进行空间目标的姿态测定.

A new method for space object pose determination by single image with known exterior orientation parameters is proposed. Initial values of object pose under world coordinate system are acquired by space resection and coordinate transformation. Then, precise pose parameters are obtained by least squares adjustment with collinearity equations where the exterior orientation parameters are known. The proposed approach avoids the synchronization problem of determining object pose with two or more cameras. Both simulated and real data experiments verify the correctness and validity of the proposed approach. The experimental results show that the precision of pose determination is significantly influenced by focal length of camera and distance between object and camera, thus the number of control points and focal length ought to be fully considered according to the precision requisition for application.