基于HMM的机载UWB SAR图像中的点目标检测  

Point Target Detection in Airborne UWB SAR Image Based on HMM

作　　者：方学立[1] 梁甸农[1] 王岩[1] 

机构地区：[1]国防科技大学电子科学与工程学院综合电子信息研究所,长沙410073

出　　处：《现代雷达》2004年第9期46-49,共4页Modern Radar

基　　金："十五"国防预研基金资助

摘　　要：UWBSAR图像中的目标检测通常指二面角目标检测 ,点目标检测由于没有发现军用车辆等人造目标这一直接效益而受到较少关注。为满足一些UWBSAR图像中检测点目标的需求 (如图像融合时的配准 ,通过路灯和树木进行道路检测和为了抑制树干杂波而识别树干杂波等 )而研究了其中的点目标检测问题。通过综合运用方向依赖滤波、隐马尔可夫模型和恒虚警率检测技术 。Normally, target detection in UWB SAR imagery refers to dihedral target detection, and point target detection has not been studied because man-made targets such as military vehicles are modeled as dihedral targets. For some particular applications (e.g. registration in image fusion, detection road via street lamps or trees, trunk clutter suppression via trunk recognition and etc), point target detection in UWB SAR image is needed and studied in this paper. Comprehensively using an aspect dependent filter, HMM and CFAR processing, a point target detection method for airborne UWB SAR is suggested in this paper. The experiment results show that the method is effective.

关 键 词：HMM 超宽带合成孔径雷达 隐马尔科夫模型 点目标检测 

分 类 号：TN957[电子电信—信号与信息处理]