基于滑动子空间的杂波样本筛选方法

Secondary Clutter Sample Selection Based on Sliding-window Sub-space Method

针对受目标污染的非均匀杂波辅助样本筛选问题,本文利用超低旁瓣子空间平滑加窗处理,克服已有的非均匀杂波样本筛选方法面临的初始协方差矩阵包含目标信号而引起的目标相消问题。一方面,采用超低旁瓣子孔径滑窗方法大幅提升旁瓣区信杂噪比,结合拟合方差偏离量提取目标大信号能量并予以剔除。另一方面,结合广义内积(Generalised Inner Product,GIP)等算法在子空间平滑的基础上进一步剔除非均匀杂波样本,从而得到独立同分布的杂波协方差矩阵。基于子孔径滑窗的杂波样本剔除方法能够大幅降低系统运算复杂度,提高系统应用效能。通过对机载实测数据的验证分析,该算法能够改善目标的输出信杂噪比,保留微动效应的离散弱目标信号,具备很高的工程应用价值。

In order to remove the target contaminated sample and select homogeneous secondary clutter training date,a ultra-low,side-lobe,sliding-window and sub-space data processing combined with GIP method is proposed.We can eliminate the contaminated sample to overcome the target self-cancellation,and eventually acquire the independent identical distributed clutter covariance.Firstly,the signal to clutter and noise ratio in side-lobe region can be dramatically increased with the sliding-window,and the signal-contaminated sample is eliminated according to the bias with the polynomial variance.Secondly,combined with the GIP method,the inhomogeneous clutter sample can be further removed on the basis of pre-processed data.The proposed method can decrease the computational complexity of secondary sample selection obviously so as to improve the system efficiency.Through the analysis and validation of the airborne real data,the signal to clutter and noise ratio,especially for weak target,can be increased by a large margin,which is of great engineering value.