James Mason, Lockheed Martin Space; Matthew Bold, Lockheed Martin Space Systems Company; Jacob Wirth, Lockheed Martin Space
Keywords: Optical Sensing, Space Domain Awareness, Multi-Spectral Instruments, Non-Resolved Object Characterization
Abstract:
Current electro-optical sensors used for Space Domain Awareness (SDA) predominantly provide two types of data: angles data for orbit determiniation, and light curve data for characterization. While the light curve data has been very useful in establishing patterns of life and stability, there is more information which can be extracted from the available electro-optical signature. The utility of such higher degree of characterization includes a more detailed understanding of signature, change detection, as well as positive ID to resolve cross-tagging or re-establishing custody after track loss.
Prior studies into the utility of spectral data for characterization have demonstrated one of the fundamental issues of collecting spectral data satellite optical signatures are often photon-starved, and the act of spreading the solar photons reflected from the satellite into a large number of spectral bins simply exacerbates the problem, reducing the signal to noise substantially.
Electro-optical sensors for SDA have shown their value when proliferated geographically at relatively low cost. While the signal to noise problem can be addressed by larger apertures, this diminishes the value of such sensors by driving cost higher. To recover signal to noise with affordable apertures, it becomes necessary to reduce the spectral resolution. This of course reduces the ability to resolve any fine spectral features.
We have incorporated a low-resolution spectrograph into one of the satellite tracking systems operated by the Lockheed Martins Advanced Technology Center. The goal is to investigate the utility of lower resolution spectra for the purposes of positive ID and change detection. The spectrograph is constructed using a transmission grating at the exit pupil of the telescope optical system, right before the final imaging lens. This has the added value that the resulting imagery contains not only the spectra of the satellite from the 1st order diffraction of the grating, but metric track data as well by using the 0th order spot that passes through the grating.
The image processing and calibration is critical to retrieving useful spectra, and must undergo a slightly different process than when processing ordinary track data for astro- and photometrics. The effect of a transmission grating is to generate both the 0th order image as well as the shifted, spectrally dispersed 1st order spectral image of the 0th order image summed in the same array. Because it contains both images, normal flat fielding to correct for vignetting/non-uniformity would incorrectly re-scale the spectra. Instead, the flat field image needs to be fit to the background data and subtracted in order to preserve the spectral data.
Since the satellite is sun-illuminated, we calibrate the spectra by observing sun-like (G2V) stars at varying elevations, then interpolating between these reference spectra to the elevations of the satellite observations. This fit will account, to some extent, for spectral extinction of the atmosphere dividing the processed satellite spectra by the sun-like reference spectra results in a reasonable estimation of the actual satellite color.
Results are presented of a small number of disparate and similar classes of LEO satellites, and the ability to provide discrimination between disparate classes, as well as the ability to classify similar classes of satellites using spectra is assessed.
© 2023 Lockheed Martin Corporation. All Rights Reserved
Date of Conference: September 19-22, 2023
Track: SDA Systems & Instrumentation