An Improved SAR Image Simulation Method for the Lunar Surface using Refined Terrain Modeling Derived from LROC Stereo Data

Yating Zhou, Yongsheng Zhou, Bo Zhu, Fei Ma, Qiang Yin, Fan Zhang

Abstract: Extensive simulation sample data are crucial for the intelligent interpretation of synthetic aperture radar (SAR) data of the lunar surface using deep learning techniques. Existing simulation methods inadequately capture the diverse features of lunar terrains, including complex craters, impact basins, and domes. This research presents an improved SAR image simulation method for the lunar surface using refined terrain modeling derived from Lunar Reconnaissance Orbiter Camera (LROC) stereo data, encompassing a wide range of lunar terrains. Firstly, morphology models for typical lunar terrains, such as simple craters, complex craters, impact basins, and domes, are established by combining fundamental principles of lunar terrain formation with diverse terrain features identified in LROC data. Then, the Advanced Two-Scale Model (ATSM) is utilized for its broad applicability in precisely calculating the scattering characteristics of the lunar surface. Finally, the SAR echo fast time-frequency domain simulation algorithm and the Chirp Scaling (CS) imaging algorithm are used to generate simulated SAR images of the lunar surface. The results demonstrate that the proposed method can simulate realistic lunar surface SAR images, offering a novel approach for constructing a comprehensive lunar surface SAR dataset and enhancing the application of deep learning in interpreting lunar surface data.