A Deep Learning-based Approach to 3D Reconstruction of Asteroids Using Ground-based Radar Data

The principal method employed for the three-dimensional(3D) reconstruction of asteroids is the optimisation method based on spherical harmonic functions. Nevertheless, this approach is inherently time-consuming and prone to local optima, which can lead to the emergence of distortions in the resulting 3D shapes. In order to address these issues, this paper puts forward a deep learning-based method for the reconstruction of asteroids in three dimensions. The dataset employed for the deep learning method comprises multi-angle range-Doppler images of asteroids consisting of a range of shapes. In order to demonstrate the robustness of this method, this paper presents the reconstruction of 3D models of asteroids with several different shapes. This paper provides a comparison of the reconstruction accuracy. The above analysis demonstrates that the proposed method is effective in reconstructing 3D models of asteroids. This paper improves the limitations of conventional methods.