Abstract
The star identification (star-ID) algorithm can match the stars captured by an optical system with a star catalog according to certain features. Star-ID has been an important research issue in many astronomical studies and a strong robust star-ID algorithm can effectively identify a certain number of stars as a standard source to correct uncalibrated telescopes. Generally, before star-ID, the celestial coordinates should be translated into the image coordinates with knowledge of optical center coordinates, image rotation angle, focal length of optical system, image sensor's pixel size and so on. For an uncalibrated telescope, the star-ID performance usually suffers from the errors or even the lack of these parameters. In this paper, a novel star-ID algorithm is devised which is based on image normalization technique and the Zernike moment such that the invariant features of asterisms are extracted instead of traditional ways. And three real images which captured via an uncalibrated ground-based telescope are used to validate our method, and the results show that it can effectively identify stars with a success rate of 99.27%, which demonstrate the robustness and accuracy of the proposed method.
Highlights
The star identification (star-ID) algorithm is a basic algorithm in the astronomical research which has a wide application in many fields such as navigation, attitude determination and error correction
The initial parameters often contain large errors which will affect the results of coordinate transformation and feature extraction, and cause a negative impact on star-ID
To solve this problem thoroughly, an extreme case is assumed where the first four parameters in Table 1 are completely missing, at the same time, there is an error in the telescope pointing, the image sensor is tilted, and the optical system distortion is unsolved
Summary
The star-ID algorithm is a basic algorithm in the astronomical research which has a wide application in many fields such as navigation, attitude determination and error correction. In 2008, Yang et al improved the grid algorithm which can extract scale-invariant features by using the distance between nearest neighboring star and the reference star as a standard scale factor [4]. By using this method, the recognition problem under translation, rotation and scale transformation is solved. The above researches have shown considerable success in star-ID, they can only partially meet the requirements under our situation To solve this problem thoroughly, an extreme case is assumed where the first four parameters in Table 1 are completely missing, at the same time, there is an error in the telescope pointing, the image sensor is tilted, and the optical system distortion is unsolved.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
