Calum Meredith, Defence Science and Technology Laboratory; Grant Privett, Defence Science and Technology Laboratory; Simon George, Defence Science and Technology Laboratory; William Feline, Defence Science and Technology Laboratory
Keywords: SSA/SDA, observation campaign, tracking, orbit determination, astronomical sensors
Abstract:
The costs of establishing multiple longitudinally-dispersed Space Domain Awareness (SDA) systems can be high. In order to address this, the use of modest, relatively inexpensive “amateur-class” astronomical sensors is being considered. There is a need to understand the trade-off required to provide more affordable SDA capability by exploiting the ever increasing range of consumer level optical imaging and sensors now available. Smaller, lower-cost optics and sensors can be versatile and easily-distributable within SDA tracking networks focussed on surveillance and tracking of Geostationary Earth Orbit (GEO). However, many have limitations in terms of field-of-view (FOV), non-optimal spatial resolution or poor tracking accuracy. For instruments with a narrow FOV, the automatic derivation of astrometric and calibrated photometric data can be non-trivial due to the low number of background stars in each frame. Examining potential techniques to overcome their limitations for GEO SDA has been the focus of a recent study aligned with the PHANTOM ECHOES experiment [1].
In this study, we considered a problem where angular resolution was deemed more important than FOV in fixed-stare GEO tracking using amateur-class commercial off-the-shelf (COTS) equipment, which led to issues in plate-solving the resulting imagery. We outline a novel image alignment technique to enable accurate astrometry to be derived from a sequence of images that, individually, could not be plate-solved.
During February and March 2021, the Mission Extension Vehicle-2 (MEV-2) conducted rendezvous and docking with Intelsat 10-02 in GEO, and was the subject of a coordinated experiment, known as PHANTOM ECHOES 2, which was reported at AMOS 2021 [1]. During this experiment, the UK Defence Science and Technology Laboratory (Dstl) undertook an initiative to acquire data from UK amateur astronomers via a collaboration with Basingstoke Astronomical Society under a continuation of the ARGUS initiative [2]. The collected dataset comprised of numerous, long image sequences (a total of 126 GB of data; typically comprising of more than 1000 frames per night from each telescope) taken with small FOV sensors with good spatial resolutions. To ensure that Intelsat 10-02 remained within frame, observations were undertaken without sidereal tracking, which caused the background stars to appear as streak/trails. When analysed, it was found that relative motion and photometry could be readily derived, but absolute values for the positions and the visual magnitudes of the target satellites could not be determined. As a result, a processing technique was developed and applied to this dataset to acquire this information by considering the observations as part of a time series.
Frequently, the same trailed background star(s) would appear in consecutive frames. Using these matching features, a technique called phase correlation was used to estimate the pixel distance offset between two or more frames. Using this pixel offset distance, these frames can be aligned and co-added to create a frame mosaic covering a larger area of the sky, until a sufficient number of star trails are present to enable plate-solving. Pixel positions of the background stars are then extracted using a traditional clumping approach (though matched filtering and other techniques would be equally applicable) and supplied to an Astrometry.net installation to plate-solve the pixel space of the entire extended mosaic and derive the associated World Coordinate System (WCS) transformations.
It was found experimentally that the technique required relaxation of the default pixel distance matching constraints within Astrometry.net; due partly to residual errors in star position matching following the phase correlation approach, and field distortion present at the field edges. A beneficial side effect of this technique was that, in areas of image overlap, the noise floor of the data is reduced, making it possible to detect fainter star trails.
This new technique enabled the Dstl team to obtain absolute astrometric data on Intelsat 10-02 and MEV2 during close-proximity manoeuvring through conversion of simple (x, y) pixel space coordinates into WCS. In addition enabled the generation of Gaia g equivalent determinations of photometric magnitude. While both results were reduced in accuracy from what might have been acquired by using a wide FOV sensor, the results were easily sufficient for tracking of GEO objects in proximity. This approach therefore allowed relatively inexpensive, small FOV, COTS equipment to contribute accurate astrometric and photometric data to the PHANTOM ECHOES experiment.
[1] George, S., Agathanggelou, A., Privett, G., Halpin, P., Feline, W., Ash, A., Chote, P., Scott, L., Skuljan, J., Alvino, A., Frith, J., PHANTOM ECHOES 2: A Five-Eyes SDA Experiment on GEO Proximity Operations, AMOS 2021 Conference (14-17 September 2021, Maui HI/Virtual), 2021.
[2] Feline, W., Privett, G., Ash, A., Harwood, N., Symons, J., Davis, J., Gainey, T., Gray, R., Lorrain, A., Murphy, J., Shave-Wall, D., Trevan, B., Wright, D., Argus: A UK Citizen Science Study in Support of the Surveillance of Space, AMOS 2019 Conference (17-20 September 2019, Maui HI, USA), 2019.
Date of Conference: September 27-20, 2022
Track: Optical Systems & Instrumentation