尝试将目标分解算法用于高分辨率极化目标的识别。基于 Krogager目标分解法 ,将各距离分辨单元的极化散射矩阵分解为 3个简单散射体的散射矩阵 ,分别代表球形散射体、二面角散射体和螺旋体散射体的散射机理。并利用求得的 3个幅度参量随径向距离的变化波形 ,提取反映目标散射中心位置和物理结构特性的特征。进一步采用 Fisher可分性准则作判据 ,从原始特征集中选取最佳特征组。对 5种飞机模型的识别结果表明 ,此基于矩阵分解的识别法具有较高的识别率 Try to use the target decomposition algorithm for high-resolution polarization target recognition. Based on the Krogager target decomposition method, the polarization scattering matrix of each distance resolving unit is decomposed into scattering matrixes of three simple scatterers, which respectively represent the scattering mechanism of spherical scatterers, dihedral scatterers and spiral scatterers. The characteristics of the target scattering center and physical structure are extracted by using the obtained three amplitude parameters with the variation of the radial distance. The Fisher divisibility criterion is further used as a criterion to select the best feature set from the original feature set. The recognition results of the five aircraft models show that the method based on matrix factorization has high recognition rate