Research On Identification Algorithm Of Weak And Small Target In Space Based On The Principle Of Star Tracker

Identification of the weak and small target in space is one of the most important tasks that spaceships need to carry out.Under complex starry background,weak and small target is little with the image size,which emitting characteristics similar to stars,mixes in the vast stars and then is unable to identify.With the mature technology of star tracker,a new method about automatic target recognition combined the principle of the Grid algorithm on identification is studied in this thesis.The thesis starts from structure and principles of star tracker,presents main algorithms for star identification and methods for target recognition.On the basis of this,the realization of the small and weak target recognition method based on the principle of star tracker is discussed.The main work focuses on the improvement of the identification method on the Grid algorithm and the feasibility of the new method.Specific content is as follows:1.A comprehensive review of the development and research status of star tracker and target recognition technology has been given,and the difficulties in the identification of small and weak target in space have also been pointed out.According to these difficulties,the research status of star identification and target recognition in the background of stars have been carried out.2.Basic research on the principles of star tracker and the methods of target recognition is presented.Starting from the construction of star tracker,the paper introduce the slits star tracker and CCD/CMOS detector star tracker,and displays the double star tracker we designed ourselves.The principles of star tracker include three aspects,star extraction,star identification and errors of star tracker.Star identification is the most important principles which contains three classical methods.And these three algorithms have their own advantages and disadvantages.Through the comparison,we can know that the grid method is the most suitable for the research work of this thesis,but it needs to be further improved.In the aspect of target recognition,this paper introduces common methods of video stream images sequence processing and the methods of target recognition in the background of stars,and the new optical correlator is also be used to identify the target preliminarily.3.To identify star maps Correctly is the basis of recognition on weak and small target.Based on the principle of the traditional Grid algorithm,the process identification on star maps is analyzed,which tells us the three reasons that the star map can not be identified: the threshold,the boundary conditions and the overlap of the projected elements.The improvements are to adjust the decision threshold of the matching number of stars and the distance to boundary dynamically and to allow more stars to join in the process of star pattern recognition;ranking all navigation stars in thestars library and numbered them according to their magnitudes,and projecting the numbers of false-recognition stars into grids,then removing false-recognition stars by counting frequency of all the stars.Through those improvements,the rate of star maps identification is increased from 95% in the traditional grid algorithm to 99%.4.To realize the method of recognition on small and weak target in space based on the principle of star tracker,the principles of star extraction and identification of star maps are presented.Most of stars in the star map can be identified correctly by the improved Grid algorithm,and the rest of the omissive stars and the target can be identified through the method of angular distance.After many experiments,the conditions and results of the identification process are analyzed,and the united method of automatic recognition of small and weak target under complex background of stars is discussed.Through research of the thesis,the Grid algorithm of star tracker is improved,and one new method of recognition on small and weak target in space based on the principle of star tracker using this algorithm is proposed,and its feasibility is verified.It provides theoretical support for target recognition in the background of stars,and has become an important guidance for the recognition of weak and small target in space.