摘要
本文提出了基于局部字典块稀疏表示的合成孔径雷达(Synthetic Aperture Radar,SAR)目标识别算法。采用各个训练类别分别对待识别样本进行重构。通过对比各个类别对于待识别样本的绝对描述能力,即重构误差的大小,确定目标类别。考虑到SAR图像方位角的敏感性,当训练样本按照方位角次序排列时,基于局部字典求解的稀疏表示系数具有块结构。为此,本文采用块稀疏算法求解局部字典上的线性表示系数从而获得更高的精度。为了验证提出算法的有效性,基于MSTAR数据集分别在标准操作条件和多种扩展操作条件下进行了目标识别实验。
This paper proposes a synthetic aperture radar(SAR)target recognition method based on block sparse representation on local dictionary.The test sample is represented by training samples from individual classes and then the absolute representation capabilities,i.e.the reconstruction errors,of different classes are compared to determine the target type.Considering the azimuthal sensitivity of SAR images,the sparse coefficient vector over the local dictionary has the block structure when the training samples are sorted according to their azimuths.Hence,the block sparse representation is employed to solve the representation coefficients on the local dictionary thus achieving higher precision.To validate the effeteness of the proposed method,experiments are conducted on public MSTAR(moving and stationary target acquisition and recognition)dataset under standard operating condition(SOC)and several extended operating condition(EOC).
作者
郭敦
吴志军
GUO Dun;WU Zhi-jun(Jiangxi Institute of Fashion Technology,Jiangxi Nanchang330201,China;Jiangxi Normal University,Jiangxi Nanchang330201,China;Hunan Institute of Science and Technology,Hunan Yueyang414006,China)
出处
《中国电子科学研究院学报》
北大核心
2019年第8期813-817,829,共6页
Journal of China Academy of Electronics and Information Technology
关键词
合成孔径雷达
目标识别
局部字典
块稀疏表示
Synthetic Aperture Radar(SAR)
target recognition
local dictionary
block sparse representation
作者简介
郭敦(1978—),男,江西人,硕士,讲师,主要研究方向为计算机应用技术,移动互联网、高校智慧校园,E-mail:lotus_summer117@163.com;吴志军(1984—),男,湖南人,硕士,讲师,主要研究方向为计算机应用技术,图像处理技术。
