Research On The Algorithm Of Relative Position And Pose Measurement Between Two Non-cooperative Space-crafts In Super Close Range Based On Circles

With the development of space technology,the number of spacecraft launched every year is increasing.With the frequent aerospace activities,the damaged and failure spacecrafts are produced and become the uncontrolled non-cooperative targets.They will take up a lot of resources and threaten the safety of astronauts and satellites.In order to ensure the space activities smoothly,the on-orbit service technology toward non-cooperative targets will become a serious problem that the space powers must face.However,the premise of the on-orbit service technology is the relative position and pose measurement of non-cooperative target.This paper focuses on non-cooperative targets pose measurement model based on circular features for the rendezvous and docking in super close range.The main work is as follows:1)the calibration kit of Matlab and OpenCV were used to calibrate the monocular and binocular cameras.Finally,the calibration results are analyzed and the accuracy of single and binocular calibration can meet the requirements of position and pose measurement.2)the key technologies of identification and positioning of circular features are analyzed.Aiming at the laboratory environment,this paper presents a fast recognition and high-precision localization algorithm for circular features in complex scenarios.This algorithm firstly adopts L0 norm minimization of image smoothing,and then cooperate with SVD decomposition in the ellipse quickly identify,finally use sliding gaussian fitting precision of edge location.this method is effective to solve the problems of extract the edge under complex background,low efficiency of identifying target features,and inaccurate positioning of target locations.3)Improves and expands the monocular vision measurement algorithm proposed by Marco Sabatini et al.,and combines the classical circular pose measuring geometry method,proposes a binocular vision measurement algorithm based on the characteristics of concentric circles.The algorithm can not only calculate the size of concentric circles and the distance between the docking ring and the engine nozzle,but also determine the position of the center of docking ring and engine nozzle without any artificial marking;Through discussing the solution of the attitude angle under three different positional relationships,this algorithm improves the angle adaptability of the model;using the model of concentric circles and matching the corner features gives the three-dimensional attitude angle information,which improves the reliability of the measurement.Finally,the algorithm of this model is simulated,The results show that the algorithm's attitude measurement error is less than 0.5°,and the position measurement error is less than 20 mm.Even if there is a certain installation error,the error of the attitude measurement can reach within 1°.4)A set of binocular vision pose measuring software is designed.The software is developed based on opencv on VS platform,and the functions and implementation methods of each module of the software are introduced in turn.Finally,the summary of the completed task is made and the further research is prospected.