与初始噪声分布无关的恒虚警处理器

为提高单元平均类恒虚警 (CA- CFAR)检测的性能 ,提出一种新的与噪声初始分布无关的恒虚警处理方法 ,即 Gauss双参数恒虚警检测 (GB- CFAR)方法。其特点是根据独立同分布噪声的非相参积累结果近似 Gauss分布的特性 ,用不同的参考单元估计积累后噪声的均值和标准差 ,然后将待检测单元减去均值估计值 ,再与标准差估计值和固定门限的乘积相比较 ,从而实现恒虚警检测。结果表明 ,由于剔除积累后的噪声均值等价于剔除了部分噪声能量 ,因此提高了信噪比 ,使得检测性能明显优于传统的 CA - CFAR检测器

CFAR processor independent of original noise distribution

A novel constant false alarm rate (CFAR) detector is presented which is independent of the original noise distribution. The Gaussian biparameteric CFAR (GB CFAR) processor improves the performance of the cell average CFAR. The processor uses the characteristic that the noncoherent integration of independent and identically distributed noise is very similar to a Gaussian distribution to obtain the mean and covariance estimaties of the integrated noise from adjacent range cells. The detection is implemented by comparing the difference between the amplitude of the resolution cell and the mean estimate with the product of the standard deviation estimate and a fixed threshold. Elimination of the mean of the integrated noise is equivalent to removing the partial noise energy which improves the signal to noise ratio (SNR) so that the detection capability is better than the traditional CA CFAR processor.