Vision Measurement Method Of Noncooperative Spacecraft And Flexible Tether In Automatic Rendezvous

Space rendezvous and docking plays an important role in the spacecraft on-orbit service,space attack and defense,and abandoned spacecraft clearance.The first step to accomplish these missions is to obtain the motion of rendezvous target.In order to reduce the dependence on the ground station and realize the space rendezvous task by on-board self-planning,tracking spacecraft should have the ability of motion measurement to rendezvous target.In the process of obtaining motion of rendezvous target based on computer vision,the image feature of target will change with the observation distance,so the vision-based measurement methods are also different under different observation distances.Especially for non-cooperative spacecraft and its flexibility structure,there are still many problems to be solved.This paper studies the image extraction algorithm and motion parameter measurement method of non-cooperative spacecraft and its flexible structure under different measurement distances during autonomous rendezvous.The major contents are as follows:For solving the problem of real space-based observation image source classification and quantity limitation,the spacecraft dynamics model and space-based optical visibility model are built in the second chapter.Based on OpenGL the simulation image of the spatial target and flexible structure under different observation distances can be obtained,which provides data sources and simulation environments for rendezvous target motion measurement.In the third chapter,the method of space target image extraction in stellar tracking mode,the orbit determination with very short arcs and precise orbit determination with cubature Kalman filter are studied respectively.For solving the multi-targets image extraction problem with large difference velocity,a maximum between-cluster variance improvement method for spatial target fringe image segmentation is proposed.This method can adaptively calculate the optimal segmentation threshold and extract multi-targets from the star background.For solving the problem of orbit determination with very short arcs,the calculation model of genetic algorithm is built for calculate initial orbit parameters of space target.The precise orbit parameters of space target are calculated with cubature Kalman filter,and simulations verify the effectiveness of the relevant algorithms.In the fourth chapter,the docking ring and solar panel is selected as feature structure for relative attitude determination,by analyzing the imaging effect of spacecraft natural features.The image extraction methods of two feature structures and the relative attitude determination method are studied.For solving the problem of binary value in the single-circle attitude determination method,a fast attitude determination method using the concentric circular structure is proposed.The normal vector of concentric circular plane is unique by using the concentric circular algebra constraint,which can determinate the relative attitude by docking ring structure with unknown center and radius.The relative attitude errors are analyzed by mathematical simulation,and relative attitude determination method by using dual feature structure is used to disabled spacecraft,which verified this method is effective for non-cooperative spacecraft relative attitude determination.In the fifth chapter,the image extraction and tracking method of the spacecraft flexible tether are studied.Tether non-uniform motion leads to uneven distribution of gray scale,low signal-to-noise ratio and motion blur on the image,so the centre line extraction algorithm and the motion blur segmentation algorithm are used in different motion speed stages,which can improve the image extraction accuracy.A tether edge segmentation operator for segmenting the motion blur region into exposure beginning and ending time tether i s designed,which not only solves the problem of low accuracy of centerline extraction algorithm under high speed motion,but also promote camera's inherent image acquisition frame rate and reduces the cost and complexity of the measurement system.A material point tracking method adapted to the bending and longitudinal elastic deformation of tether is proposed,which can track any point on tether without pasting or spraying marks,so tether dynamic characteristics are not affected during the measurement.In the sixth chapter,the physical simulation experiment platform is built for verifying the relative attitude determination method using the target spacecraft feature structure and the visual measurement method of the spacecraft tether structure.Experiments on the relative attitude determination using dual feature structures under different measurement distances,different illumination conditions and different attitude angles are designed respectively.The validity and practicability of the relative attitude determination method using dual feature structures for non-cooperative targets is verified.In the visual measurement method of tether verification experiment,the visual measurement system errors,image extraction errors and the spatial position error caused by image extraction error are analyzed,and the motion blur segmentation algorithm is verified.Using the visual measurement system with certain errors,Tether dynamics model based on the absolute node coordinate method is verified.