Vishnu Anand Muruganandan, University of Canterbury; Andrew Lambert, UNSW Canberra; Richard Clare; Richard Clarel, University of Canterbury; Steve Weddell, University of Canterbury
Keywords: Multi-exposure imaging, LEO, Deconvolution, PSF of background star, Image metrics.
Abstract:
The Boller and Chivens telescope at the University of Canterbury Mount John Observatory (UCMJO) is used for imaging satellites and space debris in Low Earth Orbit (LEO). Based on the aperture (0.61 m) and focal length (3.85 m) of the B&C, and pixel size (5.5 μm) of the camera, objects at an altitude of 400 to 1000 km can be imaged with a spatial resolution of 57 to 143 cm respectively. However, the spatial resolution of the image is reduced during the observation due to atmospheric turbulence and motion blur. Since stars are point sources at infinite distance, the perturbed Point Spread Function (PSF) of the star is useful in determining the turbulent condition of the atmosphere. Hence, LEO objects are imaged against background stars and deconvolved (post-processed) using the Lucy-Richardson (LR) algorithms to improve spatial resolution.
It is crucial to identify a star as close to, i.e., a few arcseconds, the trajectory of an LEO object as possible. If the angular distance between the star and target is too large, then the star and target do not encounter a similar volume of atmospheric turbulence. Stars brighter than an apparent magnitude of eight have high sky coverage for the identification of a natural star over the trajectory of the target. The star density is higher in the galactic plane, hence there is a high probability of finding a natural star when the target passes through this region.
During the observation, the telescope is pointed towards the natural star over which the target will transit. Since the telescope is tracking at the sidereal rate, the high orbital velocity of LEO objects leads to motion blur, and the short exposure time of 0.2 to 0.4 ms is used to compensate for motion blur. However, if the natural star is fainter than the target object, then to image the natural star a longer exposure time of 10 ms is required. To image both the natural star and satellite simultaneously at different exposure times, the incoming photon flux from the sky is split between two cameras using a beam splitter. The first camera has short exposure to image the satellite, the second camera has a long exposure to image a faint star, where both cameras are triggered simultaneously.
The proposed method is applied to image LEO satellites such as ALOS, ASTEX-1, LANDSAT-7, METEOR-M, SEASAT-1, SL-4 R/B, and ISS. These images will be deconvolved and enhancement in terms of spatial resolution will be measured using LR and image metric algorithms respectively. The Full-Width Half-Maximum (FWHM) metric is used to measure the image quality of a star. However, this method is not applicable for an extended object, such as satellites or space debris at LEO. The physical dimensions of the extended object are estimated by extracting its pixels in the major axis, minor axis, and perimeter. These image metrics are used to validate the resolution, and size of the deconvolved LEO objects.
Date of Conference: September 27-20, 2022
Track: SSA/SDA