基于预分类的全极化HRRP模型匹配目标识别方法  

Target recognition for polarimetric HRRP based on pre-classification and model matching

作　　者：吴佳妮[1] 陈永光[2] 冯德军[1] 王雪松[1] 

机构地区：[1]国防科学技术大学电子信息系统复杂电磁环境效应国家重点实验室,湖南长沙410073 [2]北京跟踪与通信技术研究所,北京100094

出　　处：《系统工程与电子技术》2016年第9期1969-1974,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(61302143;61490693;41301490);国家高技术研究发展计划(863计划)(2013AA122202)资助课题

摘　　要：全极化三维散射中心模型可准确描述目标的空间几何以及极化特征,已成为目标识别的有效手段之一。针对传统高分辨距离像的匹配算法计算量大、耗时长的不足,提出一种基于预分类的模型匹配目标识别方法,通过目标散射机理分析,对目标进行预分类,减小匹配模型数,然后利用全极化高分辨距离像的散射中心位置与极化信息构造模型匹配函数,实现了目标类别的判定。基于电磁仿真计算数据的识别实验表明,该方法具有良好的目标识别能力,相比于传统方法具有更高的识别正确率以及更低的存储量和计算量。Owing to the ability of accurately modeling geometric and polarimetric scattering feature of a target, polarimetric three-dimensional scattering center model has been prevailing as one of the effective approachesfor target recognition. With the objective of solving the high computational complexity and longtime consumptioncaused by the traditional matching method in polarimetric high resolution range profile （H R R P ） , a novelrecognition method for polarimetric H RRP based on model matching and pre-classification approaches is proposed.F irstly, the proposed method classifies the radar targets based on the scattering mechanism to reduce thenumber of matching models. Then the geometric and polarimetric information of the scattering center extractedfrom polarimetric H RRP is used to create the match indicator. Finally, the targets are recognized by employing thematch indicator. Experimental results based on electromagnetic data demonstrate that the proposed method outperformsthe traditional matching method in terms of recognition accuracy, storage consumption and computational cost.

关 键 词：全极化高分辨距离像 三维散射中心模型 模型匹配 目标识别 

分 类 号：TN95[电子电信—信号与信息处理]