航天器交会对接位姿视觉测量迭代算法

An iterative method for vision-based relative pose parameters of RVD spacecrafts

针对航天器交会对接过程中利用单目视觉进行相对位姿参数确定问题,利利用由目标航天器上4个非共面设置的特征光标点和追踪航天器上单个CCD相机组成的交会对接航天器单目视觉测量系统,推导了交会对接航天器间相对位姿参数测量的迭代算法.将相对位姿参数求解问题转化为非线性优化问题,在目标航天器像平面空间内推导优化目标函数,利用Levenberg-M arquardt算法求解该非线性优化问题,并选择一解析算法以快速求解迭代初值.仿真结果表明算法的有效性和可靠性,能够满足交会对接航天器测量精度要求.

Using a monocular vision system constituted of four non-coplanar reference points in target and a CCD camera in chaser, develops the iterative method for relative pose parameters of RVD spacecrafts by representing the relative parameters using unit quaternion. The relative pose determination is transform a nonlinear optimization calculation, the merit function is developed in image space, and Levenberg - Marquardt algorithm is utilized to minimize the merit. To provide the initial values to the iteration process, a closed -form method for relative pose parameters is selected, which greatly decreased the computational efforts of the iteration process. In the last, the solution has been validated using mathematical simulation, and simulation results show that the iterative method is effective and reliable~