基于K-均值聚类的SVD杂波抑制算法

针对强地物静止杂波及慢速杂波严重环境下，慢速运动目标被淹没其中而无法有效检测的问题，本文设计了一种基于K-均值聚类的SVD杂波抑制方法。该方法对回波信号矩阵进行奇异值分解，依据回波信号特性，得到相应的奇异值谱分布，以及奇异向量的空间相关性和平均多普勒频率三个统计量特征，然后基于这些特征采用K-均值聚类算法对各奇异分量进行聚类，无需人为设定阈值参数估计杂波基，可以自适应确定杂波子空间所对应的奇异向量，最后通过正交子空间投影来抑制回波信号中的杂波成分。实验结果表明，该方法在低信杂比条件下相比于传统子空间方法，能够得到较好杂波抑制效果。

SVD Clutter Suppression Method Based on K-Means Clustering

To detect slow moving targets in strong clutter, this paper proposes a method for clutter suppression based on the combination of K-means clustering and singular value decomposition (SVD). By SVD of the echo signal, the distribution of singular value spectrum, and the features of spatial correlation and mean Doppler frequency from the singular vector can be obtained. Based on these features, the singular components are classified by the K-mean clustering algorithm. It can directly estimate the cluster eigen rank without preset threshold on parameters, and can adaptively find the singular vector corresponding to the clutter subspace. Then, by orthogonal subspace projection, the clutter components in echo signal are suppressed well. Finally, the experimental results show that the proposed detection method has obviously better detection performance on slow moving targets under low signal-to-clutter ratio environment.