Few-Shot Multistatic ISAR Target Classification Based on Multilevel Feature Reconstruction

The collaborative observation of multistatic inverse synthetic aperture radar (ISAR) can provide comprehensive information for air/space target classification. However, under limited sample conditions, the existing methods are challenged by effective information fusion and distinguishable feature extraction. To tackle these issues, a few-shot multistatic ISAR target classification method, dubbed multilevel feature reconstruction (MLFR), is proposed in this article, which comprises two key components: interstation feature reconstruction (ISFR) and interclass feature reconstruction (ICFR). Specifically, ISFR facilitates complementary information mining and redundant information suppression by multistatic feature alignment and reconstruction, while ICFR enhances interclass separability and intraclass compactness via multiclass feature interaction and reconstruction. Additionally, a specialized hybrid loss is designed to ensure the desired outputs of the related modules. Experimental results on a multistatic ISAR dataset of satellite targets demonstrate that the proposed method significantly improves the few-shot classification accuracy, and ablation studies and visual analysis further validate the effectiveness of each module.