Joint Compensation for Spaceborne ISAR Imaging of High-Speed Space Targets via DA-BiMamba
Abstract
In spaceborne inverse synthetic aperture radar (ISAR) imaging of high-speed space targets, the high orbital velocities of the radar and target can result in pronounced translational motion. Meanwhile, the large relative radial velocity makes intra-pulse motion nonnegligible, causing high-resolution range profile (HRRP) distortion. More critically, the phase terms associated with intra-pulse and translational motion are coupled. However, existing methods typically compensate these two motion effects separately, leading to residual error propagation and limited compensation accuracy. To address this problem, a joint intra-pulse and translational motion compensation (TMC) method based on dual-axis bidirectional Mamba (DA-BiMamba) is proposed. First, a complex feature encoder is used to extract the amplitude and phase features of each HRRP. Then, DA-BiMamba is used to jointly model translational motion features along the slow-time dimension and HRRP distortion features along the range dimension. These features are then fused and mapped to the radial motion parameters. Finally, the compensation function is constructed to achieve joint compensation. Experiments on simulated and real data demonstrate the effectiveness and robustness of the proposed method.