Star Rapid Identification And Positioning Technology

Star Light Navigation is a kind of Celestial Navigation,it can obtain the current attitude and position of the spacecraft by processing,analysing and identifying the acquired star data.The algorithm of Star Location and Star identification is the key technology of Star Light Navigation. The accuracy and real-time of the navigation is affected by this algorithm,therefore taking a research on the high-precision location and rapid recognition of star image means to research the core technolog of Star Navigation,and that is the most important way to improve both accuracy and real-time of Star Light Navigation.After analysing the basic principles,the pros and cons of various algorithms of the star image location and identification,then comprehensively consider the accuracy and speed,finally presents an algorithm with high-precision positioning and fast recognition. The algorithm mainly includes three parts: pre-processing star image , star image location and star pattern identification.The pre-procession of the star image includes two parts:filtering and division.The first part is realized by the Wiener filter. Based on the original threshold segmentation algorithm, by analyseing the histogram of the star image and extracting the star points external ,the threshold of the star image division can be setted,therefore the division can be successfully achieved.In order to improve location precision, a subpixel locating technology is used in this paper.After analysing and comparing the two commonly used methods:centroid extraction and Gauss surface fitting,then based on the Gauss surface fitting,proposes a high-accuracy double-Gaussian curve fitting method.As for star image identification,an improved algorithm is proposed based on traditional triangle matching algorithm.This algorithm is mainly improved in the following fields:characteristic parameters,star triangle vector construction, navigation star database and matching algorithms.Those improvements have greatly improved the speed of identification and the recognition accuracy.