威布尔杂波背景下的恒虚警检测

Constant false alarm rate detection against Weibull clutter background

讨论了线性检波下威布尔杂波背景的恒虚警检测问题.对均匀杂波情形,由矩估计方法得到了参数估计的解析表达式,并分析了基于矩估计得到的恒虚警检测器的检测性能,结果表明:这种方法得到的估计参数的方差比由极大似然估计得到的要小,且基于此法的恒虚警检测损失比基于极大似然估计和次序统计量估计的也要小.对非均匀杂波情形,提出了贝叶斯恒虚警检测器,检测性能分析表明:贝叶斯恒虚警检测器的检测效果明显优于基于矩估计的恒虚警检测器,比基于极大似然估计和次序统计量估计的检测器效果好,且样本数越小效果越明显.

Constant false alarm rate （CFAR） detection for Weibull background and linear detector is discussed. In homogeneous background, using moment estimation we get the analytical expression of the parameters, and the performance is analyzed for CFAR-detector based on moment estimation. The analysis indicates that, the obtained variance of parameter is smaller than that using maximum likelihood estimation （MI.E） and order statistics, and CFAR loss is smaller than that using the CFAR algorithm based on MI.E. In non-homogeneous background, a Bayes-CFAR detector is presented and detecting performance is also analyzed. The performance of Bayes-CFAR detector is better than that based on moment estimation, and also better than that using MI.S and order statistics. The smaller the sample data is, the better performance it will present.