A Flexible Image Processing Framework for Vision-based Navigation Using Monocular Image Sensors

摘要: On-Orbit Servicing (OOS) encompasses all operations related to servicing satellites and performing other work on-orbit, such as reduction of space debris. includes repairs, refueling, attitude control and other tasks, which may be needed put a failed satellite back into working condition. A requires accurate position orientation (pose) information about the target spacecraft. A large quantity different sensor families is available accommodate this need. However, when it comes to minimizing mass, power required for system, mostly monocular imaging sensors perform very well. A disadvantage is- comparing LIDAR sensors- that costly computations are process the data sensor. The method presented in paper addressing these problems by aiming implement three design principles; First: keep computational burden low possible. Second: utilize algorithms and choose among them, depending on situation, retrieve most stable results. Third: Stay modular and flexible. The software designed primarily utilization where- example- servicer spacecraft approaches an uncooperative client spacecraft, can not aid any way is assumed completely passive. Image processing used navigating spacecraft. In specific scenario, vital obtain distance bearing until, last few meters, six degrees of freedom known. The smaller between more pose estimates required. The here tested optimized sophisticated Rendezvous Docking Simulation facility (European Proximity Operations Simulator- EPOS 2.0) its second-generation form located at German Space Center (GSOC) Wesling, Germany. This particular simulation environment real-time capable and provides interface test system hardware closed loop configuration. results from these tests summarized well. Finally, outlook future work given, with intention providing some long-term goals is presenting snapshot ongoing, far yet completed work. Moreover, serves overview additions which improve further.