基于分数低阶归一化匹配滤波的目标检测方法  

Method of target detection based on fractional low-order normalized matched filtering

作　　者：杨荷英[1] 王首勇[1] 郑作虎[1] 

机构地区：[1]空军预警学院

出　　处：《空军预警学院学报》2014年第5期313-317,共5页Journal of Air Force Early Warning Academy

摘　　要：在非高斯相关杂波背景下,基于归一化匹配滤波检测方法的性能随着非高斯特性增强而下降.为消除非高斯性的影响,提出一种基于分数低阶归一化匹配滤波的雷达目标检测方法.首先对观测数据进行分数低阶幂运算处理,然后基于杂波分数低阶统计量给出了检测模型,该模型的杂波散斑分量协方差矩阵通过对杂波分数低阶协方差矩阵进行归一化而得到.仿真结果表明,在非高斯杂波背景下该方法的检测性能优于归一化匹配滤波方法及传统MTD方法.Under the background of non-Gaussian correlated clutter, the performance of detection methodbased on normalized matched filtering is deteriorating as non-Gaussian characteristics are rising. To eliminate theinfluence of non-Gaussian characteristics, this paper brings forward a method of target detection based onfractional low-order normalized matched filtering. In this scheme, the operational processing with fractionallow-order power is first performed for the observed data, next, the detection model given based on the low-orderstatistics of clutter fraction and finally, the covariance matrix of clutter speckle component is achieved bynormalizing the fractional low-order covariance matrix. Simulation results show that the detection performance ofthis proposed method outperforms those of the normalized match filtering and the traditional MTD ones, under thebackground of non-Gaussian clutter.

关 键 词：非高斯相关杂波 目标检测 归一化匹配滤波 ALPHA稳定分布 分数低阶统计量 

分 类 号：TN957.51[电子电信—信号与信息处理]