High-Fidelity Star Image Simulation Under Earth-Atmosphere Background Radiation

Simulating star images under earth-atmosphere background radiation is crucial for developing anti-interference algorithms for star sensors. However, existing methods face two major limitations. Atmospheric scattering modeling typically treats the entire sensor field of view as a whole, lacking pixel-level transmission path modeling. Surface reflection models still commonly rely on the Lambertian assumption. To address these issues, this paper proposes a physics-driven, high-fidelity star image simulation framework. First, by constructing a rigorous limb observation geometry, the atmospheric scattering light transmission path corresponding to each pixel is analytically resolved, achieving high spatial resolution reconstruction of atmospheric scattering radiation intensity. Second, a surface reflection radiation model based on the Bidirectional Reflectance Distribution Function (BRDF) is introduced to accurately describe the anisotropy and regional differences of surface reflection. Finally, these models are integrated with actual multi-physics processes to build a simulation system from radiative transfer to star image generation. Experimental results show that the simulated star images are highly consistent with real captured images in both visual features and statistical distribution, with an average grayscale difference of only 0.23 on 8-bit scale. This research provides high-reliability, high-precision simulation data support for the development of earth-atmosphere background radiation suppression algorithms.