融合动态极化旋转域与状态空间模型的全极化HRRP识别OA
Fully polarimetric HRRP recognition based on the fusion of dynamic polarization rotation domain and state space model
针对空间目标全极化高分辨距离像(HRRP)识别面临的空间拓扑易丢失及小样本泛化弱等问题,提出一种融合动态极化旋转域与状态空间模型(SSM)的全极化HRRP识别方法.首先,沿空间距离维依次提取极化特征构建动态极化旋转域序列,在保留空间拓扑的同时增强物理机理表达;然后,设计物理与数据并行的双流SSM,以线性复杂度高效捕获长程空间依赖.基于多类典型卫星电磁仿真数据集的实验结果表明,该方法的整体识别精度优于对比方法,实现了数据与机理的优势互补,在小样本与低信噪比条件下展现出更优的泛化能力与鲁棒性.
Aiming at the problems such as the easy loss of spatial topology and weak generalization of small samples in the fully polarimetric high-resolution range profile(HRRP)recognition of space targets,this paper pro-poses a novel recognition method fusing the dynamic polarization rotation domain and the state space model(SSM).The proposed method sequentially extracts polarimetric features along the spatial range dimension to con-struct a dynamic polarization rotation domain sequence,enhancing the expression of physical mechanisms while retaining the spatial topology,and then designs a dual-stream SSM with parallel physics and data to efficiently cap-ture long-range spatial dependencies through linear complexity.Experimental results based on electromagnetic simulation datasets of multi-typed typical satellites demonstrate that the proposed method outperforms benchmark methods in terms of overall recognition accuracy,achieving complementary advantages between data and mecha-nisms,and exhibits better generalization capability and robustness under the conditions of small samples and low signal-to-noise ratios.
成亚玲;谭爱平
湖南工业职业技术学院,长沙 410207湖南工业职业技术学院,长沙 410207
信息技术与安全科学
空间目标识别全极化高分辨距离像动态极化旋转域双流状态空间模型
space target recognitionfully polarimetric HRRPdynamic polarization rotation domaindu-al-stream state space model
《空天预警研究学报》 2026 (4)
235-240,258,7
国家自然科学基金项目(62401589)
评论