Thermal infrared characterization of spatially unresolved resident space objects: Prospects from analytical two-component modeling

TL;DR

The paper addresses remote characterization of unresolved RSOs using thermal infrared photometry by introducing a two-component greybody spectral model that captures bimodal temperature distributions arising from spacecraft geometry. It builds an instrument-agnostic framework with Poisson-noise assumptions and analyzes identifiability and parameter uncertainties using simulated, WISE-like eight-band photometry, deriving $I(\lambda, \alpha, f_C, T_{eff}, \Delta T)$ with $T_C = T_{eff} - (1 - f_C)\Delta T$ and $T_H = T_{eff} + f_C\Delta T$. Key findings show that $T_{eff}$ can be retrieved with high precision, while $\Delta T$ and $f_C$ are more challenging but become discernible for $\Delta T_0 \gtrsim 15$ K; model preference analyses indicate the two-component model is generally favored across realistic RSOs. The study highlights that higher-bandwidth observations, priors on geometry, or multi-telescope campaigns can substantially improve parameter constraints and enable monitoring of power flows between RSO components, offering a qualitative advance in space-domain awareness. The results motivate future work on extended multi-component models, spectroscopic capabilities, and incorporating realistic systematics to translate this framework into operational SDA tools.

Abstract

In this work we investigate the potential of a thermal infrared (IR) space telescope to remotely characterize the component temperatures of a satellite. With the rapid increase in the number of objects launched in recent years, the ability to detect, track, identify and determine the intent of satellites has become of increasing importance. Spectral modeling of satellites from multi-wavelength photometry in the thermal IR is a technique that has the potential to derive information about the temperature and operational status of a satellite in orbit, without the requirement to spatially resolve the target. Previous work has focused on determination of a single/effective temperature for a Resident Space Objects (RSOs) - such as satellites, asteroids, debris and rocket bodies - from remote observations, obtaining mixed results in terms of ability to classify objects. To progress, we explore a two-greybody component spectral model. Using this analytical model, we investigate which temperature characteristics may be identified from unresolved multi-wavelength photometric observations as a function of the signal-to-noise ratio, under the assumption of Poisson noise-dominated data. With this instrument-agnostic framework, we then quantify the potential of this model to discriminate between RSOs with a single temperature (e.g. natural rocks) versus human-made satellites with a chassis and deployed solar panels where significant component temperature differences exist under typical orbital configurations. Last, we comment on promising prospects of this model for applications to existing and future space telescope observations to characterize RSOs from spatially unresolved photometry

Figures (13)

• Figure 1: Trackable RSO count evolution over time by classification. The number of objects in orbit has more than doubled in the last 10 years, and about 22 of current objects have unknown classification. Extrapolating, this trend of a larger number of satellites in orbit will require investment in new technology to expand existing SDA networks. In addition, the demand for more sophisticated space surveillance to ensure the security of critical infrastructure also highlights the need to broaden the capabilities of SDA facilities. Data sourced from ESA DISCOS (Database and Information System Characterising Objects in Space; 2024).
• Figure 2: Left panel: Schematic of a SBSS system at IR wavelengths. From its vantage point in low Earth orbit, a space telescope will be able to measure temperature and power fluctuations of other satellites by means of high signal-to-noise infrared observations of their thermal signatures. Such inference would not be possible from ground-based observatories as infrared observations are affected from substantial foreground noise emitted by the molecules in the atmosphere, illustrated in the right panel.
• Figure 3: WISE space telescope bandpasses shown as a function of wavelength, with the shaded region providing the width of the bandpass and the solid line providing the band's central wavelength. The bands are shown in blue for W1 at 3.4, orange for W2 at 4.6, green for W3 at 12, and purple for W4 at 22. The maroon and black series show blackbody spectra at 330 ($\approx$57) and 270 ($\approx$-4) respectively. The WISE bands sample the peak of the distribution. Observations in similar bands from the Midcourse Space Experiment satellite have previously been used to determine the temperature of RSOs paxson_space_2008. As such, this work considers these bands' central wavelength as the baseline for investigating a more sophisticated two-component thermal model.
• Figure 4: An example of an RSO detected in WISE band passes on the 13th February, 2010 at 11:16 UTC. The circled RSO is catalog matched to be Echostar 4, NORAD: 25331, and has a SNR of 181, 430, 1203 and 206. for W1, W2, W3, and W4 respectively for a typical WISE single frame with exposure time of 8.8 in W3 and W4, and 7.7 in W1 and W2.
• Figure 5: Illustration of a two-component 'box-wing' type satellite with a hot-component of two solar panels at temperature $T_P$ each contributing a proportion of $(1 - f_B)/2$ to the effective area and cool-component of a body at temperature $T_B$ contributing the remaining proportion $f_B$ to the effective area. The IR spectrum of such a satellite is modeled by our two-component spectral model (eqn. \ref{['eqn:2bb']}) by adding the contribution of the greybody emission at temperature $T_P$ from the solar panels and the greybody emission from the body at temperature $T_B$. When illuminated by the Sun and idle $T_P$ will typically exceed $T_B$ due to their geometry.