基于EM算法和GOF的宽带分布式目标检测算法  

A Detection Method Based on EM and GOF for Wideband Spatially Distributed Target

作　　者：李涛[1] 冯大政[1] 夏宇垠[1] 

机构地区：[1]西安电子科技大学雷达信号处理国家重点实验室,陕西西安710071

出　　处：《电子学报》2010年第10期2246-2250,共5页Acta Electronica Sinica

基　　金：国防预研基金(No.9140A07020508DZ0118)

摘　　要：针对K分布杂波背景下的宽带分布式目标检测问题,提出了一种基于期望最大化算法和拟合优度检测方法的检测算法.该算法利用期望最大化方法估计杂波参数,提高了估值精度.通过结合拟合优度检测有效利用了目标回波对杂波背景的干扰,并简要分析该算法的恒虚警率性.对仿真数据与实测数据的多个仿真结果表明,与传统的非相干积累方法及基于散射点密度的广义似然相比,该算法检测性能有明显的提高.A detection algorithm for the spatially distributed target on the background of K distribution clutter is proposed.The proposed algorithm estimates the parameters of the clutter by Expectation Maximization（EM） method and improves the estimation precision.Then combined with Goodness-Of-Fit（GOF） test,it efficiently utilize the disturbance of target echo to clutter background.Furthermore,Constant False Alarm Rate（CFAR） property with respect to the unknown clutter power is proved.Finally,experimental results for simulated data and measured data illustrate that the proposed algorithm achieves a visible performance improvement comparing with conventional non-coherent integration and Scatter Density Dependent Generalized Likelihood Rate Test（SDD-GLRT）.

关 键 词：分布式目标检测 K分布杂波 期望最大化 拟合优度检测 

分 类 号：TN957.51[电子电信—信号与信息处理]