Spaceborne SAR Image Simulation Method Based on Rigorous Geometric Location Model

Spaceborne Synthetic Aperture Radar (SAR) is widely utilized due to its capabilities of working all-time, all-weather and with wide swath. With the development of artificial intelligence technology, SAR application large models have an increasing demand for training data. However, the high acquisition cost of SAR measured data makes it difficult to obtain large-volume training samples. This paper proposes a spaceborne SAR image simulation method based on rigorous geometric location model, which enables users to quickly acquire the required SAR application training images at a relatively low cost. The framework implements satellite orbit simulation through two distinct approaches, achieves geocoding via geolocation models, computes local incidence angles for each target, and simulates pixel values using backscattering models. Experimental results demonstrate that simulated SAR images generated by the proposed method exhibit high fidelity in geometric texture patterns and intensity characteristics compared to real SAR acquisitions, validating the effectiveness of the methodology.