Ellen Glad, Millennium Space Systems, A Boeing Company
Keywords: Satellite Detection, SSA, Observations, Daytime Imaging
Abstract:
Millennium Space Systems procured a 100% commercial-off-the-shelf (COTS) mobile ground observation system in 2020 with the objective to detect and track satellites in low Earth orbit (LEO) at night, and then we created a follow-on goal in 2021 to expand this capability to the daytime. We traded multiple sensors, optical telescopes, and tracking mounts to achieve our first main objective, and we have documented the important considerations of a system intended for detecting and tracking small, dim objects. Since 2020 we have captured several visible tracks of satellites in LEO (down to 1U size) and in Geostationary orbit (GEO), as well as produced light curves for a handful of them. Detecting satellites in the visible bands relies on reflected sunlight unless the satellite produces its own radiance. To receive enough reflected solar signal from a satellite onto a visible sensor, GEO satellites can only be observed at night, and LEO satellites have an even smaller window, only being visible at twilight hours when the sun is still near the horizon.
This work discusses the expansion of our capabilities into daytime satellite detection using COTS hardware. The solar radiance spectrum extends beyond the visible into the shortwave infrared (SWIR) bands and further, so limiting transmission of a visible telescope to the SWIR decreases noise from a bright daytime background. To see if we could accomplish this goal with a COTS system, we built an optical model and traded six SWIR sensors to determine which option, when combined with either of our two fast-aperture telescopes, would produce the highest likelihood to detect satellites in a daytime background. We also simulated the use of various longpass filters to procure alongside the SWIR sensor to further reduce the daytime background signal.
Our optical modeling trades described in this paper led our team to procure the Raptor Photonics Owl 1280 InGaAs sensor and a 1500nm longpass filter. The modeling results for the Owl 1280 paired with the longpass filter showed detection of a LEO satellite in the daytime may be feasible. In our initial testing, we successfully detected stars after switching from our VIS to SWIR sensor. Out of dozens of attempts, we were able to detect and track about 10 satellites total in the morning hours as late as ~three hours post-sunrise. All successful daytime detections were of satellites in LEO, and the targets included rocket bodies, other large satellites, and the International Space Station (ISS).
Following our observations, we discuss our findings in verifying optical modeling results for both VIS and SWIR satellite detection. We also examine data collected from known stars to provide more reference points for model validation. In the VIS bands, our SNR estimates for a LEO satellite were high relative to what we observed, but our estimates for GEO altitudes were more consistent with our data. Our optical model used for daytime satellite detection set us up well for SWIR sensor selection, but future challenges remain in detecting CubeSats in LEO and satellites in GEO. The knowledge we gain from expansion of our capabilities to the daytime is an excellent tool for efficiency in future detection and tracking systems.
Date of Conference: September 27-20, 2022
Track: Optical Systems & Instrumentation