Optimized Asteroid Rotation and Shape Modeling through an Integrated Shape-from-silhouette and Optical Flow Approach

Abstract Asteroids offer valuable insights into the formation and evolution of the solar system. Understanding their rotation and shape is essential for studying asteroid collision histories and geological processes, as well as for planning exploration missions. The conventional shape-from-silhouette (SFS) method reconstructs an asteroid’s rotation and shape using multiple resolved silhouette images captured at different rotational states. However, this method relies solely on silhouette information and is affected by the inherent shape symmetries, leading to ambiguities in rotation and shape model inversion. This paper presents an approach that integrates SFS and optical flow techniques for optimized asteroid rotation and shape modeling. In our approach, an optical flow method revealing information on asteroid motion and structure is integrated with the SFS technique to realize effective optimization. Additionally, a swarm intelligence-based strategy relying on parallel exploration and dynamic step adjustment is used to enhance global optimization. Experiments based on data from two typical asteroids, Itokawa (asymmetric shape) and Bennu (symmetric shape), demonstrate the promising performance of the proposed method. The results show that the approach achieves unique and accurate pole estimations from distances of several hundred kilometers, even under varying illumination conditions. Additionally, it can generate unique and precise shape models when favorable illumination conditions are present. For the asymmetric asteroid Itokawa, the maximum distance that allows for pole and shape inversion is approximately twice that of the symmetric asteroid Bennu. The proposed method can support scientific research on asteroids, such as their dynamics and morphology, as well as asteroid exploration missions.