杨磊^,,
李埔丞,
李慧娟,
方澄
中国民航大学天津市智能信号与图像处理重点实验室 天津 300300
基金项目:国家自然科学基金(61601470),天津市自然科学基金(16JCYBJC41200),中央高校基本科研业务费专项资金(3122018C005)

详细信息

作者简介:杨磊：男，1984年生，副教授，研究方向为高分辨SAR成像及机器学习理论应用
李埔丞：男，1992年生，硕士生，研究方向为高分辨SAR成像稀疏特征增强
李慧娟：女，1996年生，硕士生，研究方向为高分辨SAR成像稀疏特征增强
方澄：男，1980年生，讲师，研究方向为深度学习及高性能计算

通讯作者:杨磊　yanglei840626@163.com

中图分类号:TN957.52

计量

文章访问数:2216
HTML全文浏览量:898
PDF下载量:69
被引次数:0

出版历程

收稿日期:2019-03-22
修回日期:2019-08-23
网络出版日期:2019-09-12
刊出日期:2019-12-01

Robust and Efficient Sparse-feature Enhancementfor Generalized SAR Imagery

Lei YANG^,,
Pucheng LI,
Huijuan LI,
Cheng FANG
Tianjin Key Laboratory for Advanced Signal Processing, Civil AviationUniversity of China, Tianjin 300300, China
Funds:The National Natural Science Foundation of China (61601470), The Natural Science Foundation of Tianjin, China (16JCYBJC41200), The Fundamental Research Funds for the Central Universities of Ministry of Education of China (3122018C005)


摘要
摘要:针对合成孔径雷达(SAR)成像中的稀疏特征增强问题，传统方法难以在精度与效率之间实现有效的平衡。该文提出基于复数交替方向多乘子方法(C-ADMM)，针对SAR稀疏特征增强建立增广的拉格朗日优化方程，并引入复数${\ell _1}$范数邻近算子，基于高斯-赛德尔思想进行对偶迭代运算，从而在复数回波数据域内对多种SAR模式的实测数据进行成像。实验部分首先通过仿真数据的相变图(PTD)验证C-ADMM算法对于复数数据的稀疏恢复性能，然后选取地面静止场景和地面运动目标的原始SAR图像和逆SAR图像实测数据，与凸优化(CVX)方法和贝叶斯压缩感知(BCS)方法进行对比试验，最后验证了该文所提算法在稀疏特征增强应用中的稳健性、高效性和通用性。
关键词:合成孔径雷达/
稀疏特征增强/
复数交替方向多乘子方法/
增广拉格朗日优化方程
Abstract:For the problem of sparse feature enhancement in Synthetic Aperture Radar (SAR) imagery, conventional methods are difficult to achieve a preferable balance between accuracy and efficiency. In this paper, a robust and efficient SAR imaging algorithm based on Complex Alternating Direction Method of Multipliers(C-ADMM) is proposed for general SAR imaging feature enhancement within complex raw data domain. The problem is firstly imposed by an augmented Lagrange function, and the complex ${\ell _1}$-norm of the intended SAR image is jointly formulated within the C-ADMM framework. Then, the proximal mapping of the sparse feature is derived as a soft-thresholding operator. Further, an iterative processing procedure is designed according to Gaussian-Deidel principle, and the convergence of the proposed algorithm is analyzed. In the experiment, the performance of the proposed algorithm is firstly examined by the simulated data in terms of Phase Transition Diagram (PTD) under different under-sampling rate and degree of sparsity. Then, various raw SAR and Inverse SAR(ISAR) data, for both stationary ground scene and Ground Moving Target Imaging(CMTIm), are applied to further verifying the proposed C-ADMM, and comparisons with classical Convex(CVX) and Bayesian Compress Sensing(BCS) algorithms are performed, so that both the effectiveness and superiority of the C-ADMM algorithm can be verified.
Key words:Synthetic Aperture Radar (SAR)/
Sparse feature enhancement/
Complex Alternating Direction Method of Multipliers (C-ADMM)/
Augmented Lagrangian function



PDF全文下载地址:

https://jeit.ac.cn/article/exportPdf?id=47a96832-8d4b-4502-89b1-939e35b39e23