André Gaudin, University of Canterbury; Richard Clare, University of Canterbury; Vishnu Anand Muruganandan, University of Canterbury; Steve Weddell, University of Canterbury
Keywords: SSA, Point spread function, PSF, Blob detection, natural guide star search, satellite imaging
Abstract:
Due to the large number of unresponsive satellites and space debris it can be dangerous to launch additional satellites due to the increased likelihood of collisions. It has therefore become increasingly important to identify these objects to minimise the risk to functioning satellites. Telescopic imaging of artificial satellites through the Earth’s atmosphere will typically result in images with faint or distorted outlines of objects. To reduce this distortion, either artificial or natural beacons can be used. The latter method uses the Point Spread Function (PSF) of the natural stars in the background to restore the image of artificial satellites in the foreground. This paper considers the Laplacian of Gaussian (LoG) and Difference of Gaussian (DoG) blob detection algorithms to perform natural guide star (NGS) searches over a wide field-of-view (FoV). The results of this research will be used to determine one or more PSFs, representing natural background stars. Results from a comparison of these algorithms, used to detect k NGS’s over a wide FoV from field images is presented. The algorithms detect both focused and defocused stars around localised regions, where artificial satellites in low-Earth orbit (LEO) can be detected. Search patterns are performed to support multiple locations on the trajectory of artificial satellites. Particular emphasis is placed on the computational efficiency of the algorithms to achieve real-time updates. It is expected that this method will improve the results from images of satellites by employing a fast and accurate detection algorithm for classification of natural source beacons.
Date of Conference: September 14-17, 2021
Track: Non-Resolved Object Characterization