Simulation method for all-time polarization star sensor star maps in short-wave infrared near-Earth space

The polarization star sensor is a novel concept developed to mitigate the intense background radiation in near-Earth space. To address the challenges of high costs associated with acquiring real-world images and the scarcity of effective datasets for performance analysis and algorithm training, this paper proposes a simulation method for generating polarization star sensor star maps. This method is built upon directional calculations, energy conversion, image quality degradation, and star map dataset generation. A key innovation of this method is the first-time fusion of polarization information in star map simulation. It is built upon the following key components: a short-wave infrared (SWIR) all-sky light field vector simulation algorithm for rapid retrieval of virtual scene data, a polarization filter calculation model, and a detector noise calculation model leveraging SWIR subwavelength metallic grating properties. The final output is a dataset of polarization-filtered, degree of polarization, and accumulated star maps. The validity and accuracy of the sky radiance model, the sky polarization model, and the polarization star sensor simulation results were comprehensively evaluated using theoretical, simulated, and measured data. The simulated star maps exhibit Signal-to-Noise Ratio Enhancement(SNE) errors of less than 4.88% against theoretical analysis and 14.74% against measured star maps. Comparative analysis demonstrates a marked efficiency advantage of the proposed method for multi-field-of-view or sequential star map simulation. The findings provide a key theoretical basis for informing the design, performance assessment, and algorithm development of polarization star sensors.