基于非参估计的局部滑窗双参CFAR目标检测

摘要

针对特定杂波概率模型不能有效的描述SAR图像背景杂波这一问题,提出了一种基于非参估计的局部滑窗双参CFAR目标检测算法.该方法首先用非参估计方法逼近SAR图像局部背景,完成对局部背景的精确建模；在此基础上,理论推导了局部双参CFAR检测算法的阈值,设计了阈值求解的数值算法.对典型目标图像进行实验,结果表明,该方法检测速度较快、精度较高.%For certain probabilistic model cant describe SAR image's background effectively, a target detecting algorithm of local gliding window two parameter CFARs based on non-parameter estimation is proposed. The method of non-parameter estimation is used to approximate the local background and complete the accurate model of SAR image. The threshold of local two-parameter CFAR is deduced theoretically , and numerical solution of it is also designed. Experiment of typical target images is carried out to show the rapidity and precision of the algorithm.