Visual Navigation And Tracking Study Of Non-cooperative Target By Space Manipulator

At present, Capturing technology of non-cooperative target has been a new development trend in the field of space manipulator technique, aiming at on-orbit service of the traditional spacecraft and eliminating the space junk. Visual navigation and tracking technology of the non-cooperative target,as a key technology in the process of capturing,is the research focus of scholars all over the world. In order to solve the visual navigation and tracking problem of non-cooperative target, this paper studies pose measurement and target tracking technology.Firstly, stereo vision measurement model based on the position of binocular is set up.Then the tracking method of moving target is studied, finally the validity of model and method is verified through the simulation experiment. The main content is as follows:1) This paper processes 2-D image, and puts forward the non-cooperative target recognition method based on the characteristics of information fusion. As the non-cooperative target cannot provide effective cooperation information, this method get the basic shape of rectangular and circular feature information with the characteristic of the spacecraft’s own information as object recognition, to complete the target recognition and plane center finding. Then the feature points obtained from left and right cameras are matched according to the scale invariant feature transform algorithm.2) A method based on binocular vision is proposed to measure the relative pose between the two non-cooperative spacecrafts. The scheme chooses the center of target satellite as the origin to establish non-cooperative target coordinate system, and to compute the relative position. Then the stereoscopic depth and parallax relation model is set up to analyze the error of position calculation. Finally, modeling and simulation with OpenGL is carried out to verify the feasibility, the accuracy and real-time performance of the proposed scheme.3) Visual tracking of non-cooperative target is implemented. In response to the nonlinear high maneuvering target tracking problem, the IMM algorithm is introduced into this article based on Kalman filter to enhance the performance of speed tracking. Then the performances of the two kinds of filter algorithm are compared in three different motion models. The three kinds of models are respectively the model with constant velocity(CV)and constant acceleration plants(CA) and complex curve model.Experiment analysis shows that the proposed non-cooperative target recognition method is effective to ensure the accuracy of target recognition. In the same time, pose measurement scheme and the IMM filter tracking method are accurate and easy.Improvement of the model and optimization of the algorithm is useful to significantly improve the non-cooperative target spacecraft visual navigation tracking precision.