3-D Reconstruction of Space Target Based on Silhouettes Fused ISAR–Optical Images
Bo Long, Pengling Tang, Feng Wang, Ya‐Qiu Jin
Abstract
The space situational awareness (SSA) is a critical issue of space security. The SSA of three-dimensional (3-D) information on flying space targets has been operated by inverse-SAR (ISAR) and telescope observations. However, how good it is to fuse these two observations, i.e. microwave and optical, has not been well discussed. Especially, it would be difficult to determine the dynamic attitude and its projections of a spinning target, only based on ISAR. This paper presents a co-location fusion of ISAR-optical images to acquire a 3-D reconstruction. Making use of semantic segmentation and parallelogram fitting, the dynamic parameters of the target can be acquired, and their projections can be determined. Instead of adopting a scattering point-based method, this paper presents a uniform projection representation by fusing the joint ISAR-optical silhouette-based 3-D reconstruction. This approach enables structure-level voxel reconstruction, even for a spinning target. The simulated datasets of the orthogonal projection geometry of fused space target ISAR-optical images demonstrate that good accuracy and favorable 3-D reconstruction can be achieved.