An Adaptively Truncated Clutter-Statistics-Based Two-Parameter CFAR Detector in SAR Imagery
An Adaptively Truncated Clutter-Statistics-Based Two-Parameter CFAR Detector in SAR Imagery
复制标题
SAR 图像中基于自适应截断杂波统计的二参数恒虚警检测器
DOI：
10.1109/joe.2017.2768198
复制
发表时间：
2018-01-01
影响因子：
4.1
通讯作者：
Zhou, Fang
中科院分区：
文献类型：
--
作者：
Ai, Jiaqiu;Yang, Xuezhi;Zhou, Fang
传统的恒虚警检测器在拥挤的港口和忙碌的航线中，由于受到干扰舰船目标、旁瓣和虚警等异常值的影响，检测概率会下降。本文提出了一种新的基于自适应截断杂波统计量的双参数恒虚警检测器（TS-LNCFAR）。新的双参数恒虚警检测器采用对数正态分布作为统计模型，通过自适应阈值背景窗杂波截断，在最大程度上保留了真实的杂波的同时，去除杂波样本中的野值。通过最大似然估计，使用截断杂波统计量精确地建立对数正态模型。与传统的恒虚警检测器相比，TS-LNCFAR的参数估计精度更高，在多目标环境下具有更好的虚警调节性能和更高的PD。此外，参数估计和阈值计算不需要迭代数值计算，TS-LNCFAR具有较高的计算效率。在多视Envisat-ASAR和TerraSAR-X数据上验证了该检测器的优越性。
Traditional constant false alarm rate (CFAR) detectors suffer probability of detection (PD) degradation influenced by the outliers such as interfering ship targets, side lobes, and ghosts, especially in crowded harbors and busy shipping lines. In this paper, a new two-parameter CFAR detector based on adaptively truncated clutter statistics (TS-LNCFAR) is proposed. The new two-parameter CFAR detector uses log-normal as the statistical model; by adaptive-threshold-based clutter truncation in the background window, the outliers are removed from the clutter samples, while the real clutter is preserved to the largest degree. The log-normal model is accurately built using the truncated clutter statistics through the maximum-likelihood estimator. Compared with traditional CFAR detectors, the parameter estimation is more accurate, and TS-LNCFAR has a better false alarm regulation property and a high PD in a multiple-target environment. Furthermore, the parameter estimation and threshold calculation do not need iterative numerical calculation, and TS-LNCFAR has a high computational efficiency. The superiority of the proposed TS-LNCFAR detector is validated on the multilook Envisat-ASAR and TerraSAR-X data.