基于全变分的高分辨SAR联合特征增强成像算法
作者:
作者单位:

中国工程物理研究院 电子工程研究所 , 四川 绵阳 621999

作者简介:

通讯作者:

中图分类号:

基金项目:

装备预研基金重点项目(661406190101)


Joint feature enhancement for high resolution SAR imaging based on total variation regularization
Author:
Affiliation:

Institute of Electronic Engineering, Chinese Academy of Engineering Physics, Mianyang 621999, China

Fund Project:

Equipment Pre-research Fund (661406190101)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    稀疏约束下的合成孔径雷达(Synthetic Aperture Radar, SAR)成像技术,通过对稀疏先验建模的稀疏特征进行增强,能有效获取目标特显点的有用信息,但无法对目标的结构特征进行恢复,且对不可避免的非系统误差十分敏感。为此,提出一种依靠交替方向多乘子法(Alternating Direction Method of Multipliers,ADMM)面向结构特征增强的稀疏恢复高分辨SAR成像(Structure-feature Enhancement-ADMM,SE-ADMM)算法。该算法引入全变分(Total Variation,TV)正则项建模结构特征,起到增强结构的作用;引入范数建模稀疏特征,起到压制噪声作用;引入最小熵范数建模聚焦特征,以保证算法对非系统乘性误差的不敏感性。在ADMM多特征优化框架下,利用“局部-全局”的运算机制,首先分别进行三个特征的邻近算子推导,以获得对应特征解析解,再进行目标全局优化保证特征解之间的协调平衡,以实现目标的多特征增强。另外,ADMM多特征优化框架下变量分裂和多正则项的引入,保证了算法的效率和稳健性。实验部分先后选取SAR仿真数据与实测数据来验证算法的有效性,通过相变热力图定量分析所提算法的恢复性能,进而验证了所提SE-ADMM算法的稳健性与优越性。

    Abstract:

    Synthetic Aperture Radar (SAR) imaging under sparse constraint can effectively obtain useful information of the target''s distinctive points by enhancing the sparse features with the sparse prior representation. However, this process cannot recover the structure feature of the target, and it is very sensitive to inevitable non-systematic errors. To this end, this paper proposes a sparse recovery high-resolution SAR imaging algorithm for Structure feature Enhancement based on Alternating Direction Method of Multipliers (ADMM) method (SE-ADMM). The algorithm introduces Total Variation (TV) regular term to characterize structural features and play a role in enhancing the structure, introduces norm to represent sparse features, which can suppress noise, and the entropy norm is introduced to characterize the focusing feature to ensure that the algorithm is insensitive to non-systematic errors. Under the framework of ADMM multi-feature optimization, the "Local-Global" operation mechanism is used to first derive the proximal operators of the three features respectively to obtain the corresponding feature analytical solutions, and then perform the target global optimization to ensure the coordination and balance between the feature solutions. In addition, the reference of multi-splitting variables and multi- regular under the ADMM multi-task framework ensures the efficiency and robustness of the algorithm. In the experimental part, the simulation data and measured data of SAR are selected successively to verify the effectiveness of the algorithm. The recovery performance of the proposed algorithm was quantitatively analyzed through phase transition analysis, and the robustness and advantages of SE-ADMM algorithm proposed in this paper are further verified.

    参考文献
    相似文献
    引证文献
引用本文

黄博,周劼,江舸.基于全变分的高分辨SAR联合特征增强成像算法[J].红外与毫米波学报,2021,40(5):664~672]. HUANG Bo, ZHOU Jie, JIANG Ge. Joint feature enhancement for high resolution SAR imaging based on total variation regularization[J]. J. Infrared Millim. Waves,2021,40(5):664~672.]

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-12-27
  • 最后修改日期:2021-09-06
  • 录用日期:2021-05-18
  • 在线发布日期: 2021-09-03
  • 出版日期: