CFAR technique is widely used in radar targets detection fields. Traditional algorithm is cell averaging (CA), which can give a good detection performance in a relatively ideal environment. Recently, censoring techniq...CFAR technique is widely used in radar targets detection fields. Traditional algorithm is cell averaging (CA), which can give a good detection performance in a relatively ideal environment. Recently, censoring technique is adopted to make the detector perform robustly. Ordered statistic (OS) and trimmed mean (TM) methods are proposed. TM methods treat the reference samples which participate in clutter power estimates equally, but this processing will not realize the effective estimates of clutter power. Therefore, in this paper a quasi best weighted (QBW) order statistics algorithm is presented. In special cases, QBW reduces to CA and the censored mean level detector (CMLD).展开更多
针对距离扩展目标所占据的各距离单元输出的信杂比在统计意义上不同的一般情况,在分析基于顺序统计的广义似然比检测(generalized likelihood ratio test based on order statistics,OS-GLRT)的基础上提出了两种改进的检测器。第一种是...针对距离扩展目标所占据的各距离单元输出的信杂比在统计意义上不同的一般情况,在分析基于顺序统计的广义似然比检测(generalized likelihood ratio test based on order statistics,OS-GLRT)的基础上提出了两种改进的检测器。第一种是修正的OS-GLRT,它采用或规则,将所有可能的有效距离单元数量得到的OS-GLRT的结果进行融合后做出判决,所得到的修正的OS-GLRT无需有效距离单元数量等目标先验信息。第二种是结合双门限(dual threshold,DH)技术,通过设置固定的第一门限对有效距离单元数目进行估计,动态的第二门限形成最终的判决。与同类检测器相比,新检测器不需要距离扩展目标的先验信息,具有较好的鲁棒性。展开更多
文摘CFAR technique is widely used in radar targets detection fields. Traditional algorithm is cell averaging (CA), which can give a good detection performance in a relatively ideal environment. Recently, censoring technique is adopted to make the detector perform robustly. Ordered statistic (OS) and trimmed mean (TM) methods are proposed. TM methods treat the reference samples which participate in clutter power estimates equally, but this processing will not realize the effective estimates of clutter power. Therefore, in this paper a quasi best weighted (QBW) order statistics algorithm is presented. In special cases, QBW reduces to CA and the censored mean level detector (CMLD).
文摘针对距离扩展目标所占据的各距离单元输出的信杂比在统计意义上不同的一般情况,在分析基于顺序统计的广义似然比检测(generalized likelihood ratio test based on order statistics,OS-GLRT)的基础上提出了两种改进的检测器。第一种是修正的OS-GLRT,它采用或规则,将所有可能的有效距离单元数量得到的OS-GLRT的结果进行融合后做出判决,所得到的修正的OS-GLRT无需有效距离单元数量等目标先验信息。第二种是结合双门限(dual threshold,DH)技术,通过设置固定的第一门限对有效距离单元数目进行估计,动态的第二门限形成最终的判决。与同类检测器相比,新检测器不需要距离扩展目标的先验信息,具有较好的鲁棒性。
