宽角域合成孔径雷达（wide-angle synthetic aperture radar, WA-SAR）有着更广泛的角度覆盖范围，基于此得到的宽角域散射中心（wide-angle scattering centers, WA-SCs）包含了目标物体更加丰富的电磁散射特征，这对雷达的目标建模、目标识别等有着重要的意义。为了克服WA-SCs数据维度高、所含信息复杂的特点，并从中提取出所需的目标物体特征，采取密度峰值聚类（density peak clustering, DPC）算法研究WA-SCs。基于SLICY模型数据，从聚类内部评价指标、聚类可视化和算法自动化程度3个方面，将本文算法与经典的K-means、DBSCAN和MeanShift算法进行了对比实验。结果表明，DPC算法具有自动化程度高、高维数据适应性强、聚类精度高等优点，有望为后续的一系列基于WA-SCs的目标建模、目标识别等工作提供技术支撑。

Wide-angle synthetic aperture radar (WA-SAR) has a broader coverage of angles, and wide-angle scattering centers (WA-SCs) derived from it encompass richer electromagnetic scattering characteristics of target objects, which is important for the following analysis. To address the high-dimensional and complex nature of WA-SCs data and extract targets' features, density peak clustering (DPC) was applied to WA-SCs. Based on the SLICY model dataset, from three aspects of clustering internal evaluation, clustering visualization and automation of algorithm, DPC was compared with three classical K-means, DBSCAN and MeanShift algorithms. The results show that DPC has advantages of high degree of automation, high dimensional data adaptability, high accuracy of clustering and so on, which is expected to provide technical support for target modeling and target recognition through WA-SCs.