提出了基于立体视觉的空间机器人捕获非合作目标的测量与规划方法,包括图像处理与特征提取、3D重构及位姿测量、自主规划与控制等.同时,还建立了一套仿真系统,可进行包括图像采集、位姿测量、追踪星GNC、空间机器人规划与控制、航天器动力学计算等在内的闭环仿真.仿真结果验证了算法的有效性.
A stereo vision-based measurement and planning method is proposed for space robot to capture a noncooperative target. The method is composed of image processing and feature extraction, 3D reconstruction and pose measurement, and autonomous planning and control. Furthermore, a simulation system is created, and the closed-loop simulations, including image collection, pose measurement, chaser GNC (guide, navigation and control), space robot planning and control, and satellites' dynamic calculation, are conducted. The simulation results verify the algorithms above.