基于纹理分割恒虚警检测器设计与验证  

Design and Validation of Texture Segmentation Based-CFAR Detector

作　　者：张维[1] 于盛林[1] 张弓[2] 冯强[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《南京航空航天大学学报》2008年第1期80-84,共5页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：航空科学基金(04D52032)资助项目

摘　　要：提出一种基于纹理分割恒虚警(TSB-CFAR)检测器,旨在解决杂波边缘情况下恒虚警检测器性能急剧下降的问题。该检测器由杂波分割模块和变参考单元CFAR模块组成,前级杂波分割模块通过多重分形纹理分割算法对杂波图进行区域分割,跟踪杂波边缘在当前参考单元中的位置,后级CFAR检测模块根据杂波边缘位置,选择合适的参考单元参与目标检测。理论分析表明该检测器不仅具有良好的杂波边缘虚警控制能力,而且在均匀背景下能够达到单元平均恒虚警(CA-CFAR)检测器的性能。IPIX海杂波和MSTAR合成孔径雷达地杂波数据检测实验也证实了本文方法的有效性。A texture segmentation based-constant false alarm rate （TSB-CFAR） detector is designed to solve the clutter edge performance degradation problem for many conventional detectors. TSB-CFAR consists of a clutter map texture segmentation module and a varying reference cell CFAR detection module. The segmentation module performs multifractal segmentation to the clutter map and tracks clutter edge locations in reference cells while the posterior CFAR detection module selects proper reference cells for detection according to anterior results. Theoretical analyses show that TSB-CFAR has an excellent performance in clutter edge scenarios and maintains an equivalent performance to CA-CFAR in homogeneous scenarios. The MSTAR SAR ground TSB-CFAR detector is proved to be effective segmentation by IPIX sea clutter and clutter detection experiments.

关 键 词：恒虚警 杂波跟踪 纹理分割 多重分形分割 

分 类 号：TN958[电子电信—信号与信息处理]