摘要
针对合成孔径雷达(SAR)图像目标识别中存在物体遮挡的情况,该文提出一种基于非负稀疏表示的分类方法。通过分析L0范数和L1范数最小化在求解非负稀疏表示问题上的区别,证明在一定条件下,L1范数最小化方法除了保持解的稀疏性还能得到与输入信号更加相似的原子集合,因此也更加适用于分类问题中。在运动和静止目标获取与识别(MSTAR)数据集上的识别实验结果表明,采用L1范数的非负稀疏表示分类方法能达到较好的识别性能,并且相对传统方法对存在遮挡情况下的识别问题更稳健。
In order to solve the occlusion issue in SAR image target recognition, a new classification method is proposed based on non-negative sparse representation. The difference between L0-norm and Ll-norm minimization in solving non-negative sparse representation problem is analyzed, and it is proved that Ll-norm regularization method pursuits not only the sparsity of the solution but also the similarity between the input signal and the selected atoms under some conditions, hence it is fit for classification application. The experimental results on Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset show that the non-negative sparse representation classification method with Ll-norm regularization can achieve much better recognition performance and it is more robust in the recognition of targets with occlusion compared with the traditional method.
出处
《电子与信息学报》
EI
CSCD
北大核心
2014年第9期2194-2200,共7页
Journal of Electronics & Information Technology
基金
国家自然科学基金(61271024
61201292
61201283)
新世纪优秀人才支持计划(NCET-09-0630)
全国优秀博士学位论文作者专项资金(FANEDD-201156)
国家部委基金
中央高校基本科研业务费专项资金资助课题
关键词
SAR目标识别
非负稀疏表示
L1范数最小化
SAR target recognition
Non-negative sparse representation
Ll-norm minimization