A Novel CS Model and Its Application in Complex SAR Image Compression

Wentao LV; Gaohuan LV; Junfeng WANG; Wenxian YU

doi:10.1587/transfun.E96.A.2209

A Novel CS Model and Its Application in Complex SAR Image Compression

Wentao LV, Gaohuan LV, Junfeng WANG, Wenxian YU

Full Text Views

0

Share
Cite this

Summary :

In this paper, we consider the optimization of measurement matrix in Compressed Sensing (CS) framework. Based on the boundary constraint, we propose a novel algorithm to make the “mutual coherence” approach a lower bound. This algorithm is implemented by using an iterative strategy. In each iteration, a neighborhood interval of the maximal off-diagonal entry in the Gram matrix is scaled down with the same shrinkage factor, and then a lower mutual coherence between the measurement matrix and sparsifying matrix is obtained. After many iterations, the magnitudes of most of off-diagonal entries approach the lower bound. The proposed optimization algorithm demonstrates better performance compared with other typical optimization methods, such as t-averaged mutual coherence. In addition, the effectiveness of CS can be used for the compression of complex synthetic aperture radar (SAR) image is verified, and experimental results using simulated data and real field data corroborate this claim.

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E96-A No.11 pp.2209-2217

Publication Date: 2013/11/01

Publicized

Online ISSN: 1745-1337

DOI: 10.1587/transfun.E96.A.2209

Type of Manuscript: PAPER

Category: Digital Signal Processing

Authors

Wentao LV
  Shanghai Jiao Tong University
Gaohuan LV
  Shanghai Jiao Tong University
Junfeng WANG
  Shanghai Jiao Tong University
Wenxian YU
  Shanghai Jiao Tong University

Keyword

compressed sensing, mutual coherence, lower bound, Gram matrix, complex SAR image compression

Cite this

Copy

Wentao LV, Gaohuan LV, Junfeng WANG, Wenxian YU, "A Novel CS Model and Its Application in Complex SAR Image Compression" in IEICE TRANSACTIONS on Fundamentals, vol. E96-A, no. 11, pp. 2209-2217, November 2013, doi: 10.1587/transfun.E96.A.2209.
Abstract: In this paper, we consider the optimization of measurement matrix in Compressed Sensing (CS) framework. Based on the boundary constraint, we propose a novel algorithm to make the “mutual coherence” approach a lower bound. This algorithm is implemented by using an iterative strategy. In each iteration, a neighborhood interval of the maximal off-diagonal entry in the Gram matrix is scaled down with the same shrinkage factor, and then a lower mutual coherence between the measurement matrix and sparsifying matrix is obtained. After many iterations, the magnitudes of most of off-diagonal entries approach the lower bound. The proposed optimization algorithm demonstrates better performance compared with other typical optimization methods, such as t-averaged mutual coherence. In addition, the effectiveness of CS can be used for the compression of complex synthetic aperture radar (SAR) image is verified, and experimental results using simulated data and real field data corroborate this claim.
URL: https://globals.ieice.org/en_transactions/fundamentals/10.1587/transfun.E96.A.2209/_p

Copy

@ARTICLE{e96-a_11_2209,
author={Wentao LV, Gaohuan LV, Junfeng WANG, Wenxian YU, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={A Novel CS Model and Its Application in Complex SAR Image Compression},
year={2013},
volume={E96-A},
number={11},
pages={2209-2217},
abstract={In this paper, we consider the optimization of measurement matrix in Compressed Sensing (CS) framework. Based on the boundary constraint, we propose a novel algorithm to make the “mutual coherence” approach a lower bound. This algorithm is implemented by using an iterative strategy. In each iteration, a neighborhood interval of the maximal off-diagonal entry in the Gram matrix is scaled down with the same shrinkage factor, and then a lower mutual coherence between the measurement matrix and sparsifying matrix is obtained. After many iterations, the magnitudes of most of off-diagonal entries approach the lower bound. The proposed optimization algorithm demonstrates better performance compared with other typical optimization methods, such as t-averaged mutual coherence. In addition, the effectiveness of CS can be used for the compression of complex synthetic aperture radar (SAR) image is verified, and experimental results using simulated data and real field data corroborate this claim.},
keywords={},
doi={10.1587/transfun.E96.A.2209},
ISSN={1745-1337},
month={November},}

Copy

TY - JOUR
TI - A Novel CS Model and Its Application in Complex SAR Image Compression
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 2209
EP - 2217
AU - Wentao LV
AU - Gaohuan LV
AU - Junfeng WANG
AU - Wenxian YU
PY - 2013
DO - 10.1587/transfun.E96.A.2209
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E96-A
IS - 11
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - November 2013
AB - In this paper, we consider the optimization of measurement matrix in Compressed Sensing (CS) framework. Based on the boundary constraint, we propose a novel algorithm to make the “mutual coherence” approach a lower bound. This algorithm is implemented by using an iterative strategy. In each iteration, a neighborhood interval of the maximal off-diagonal entry in the Gram matrix is scaled down with the same shrinkage factor, and then a lower mutual coherence between the measurement matrix and sparsifying matrix is obtained. After many iterations, the magnitudes of most of off-diagonal entries approach the lower bound. The proposed optimization algorithm demonstrates better performance compared with other typical optimization methods, such as t-averaged mutual coherence. In addition, the effectiveness of CS can be used for the compression of complex synthetic aperture radar (SAR) image is verified, and experimental results using simulated data and real field data corroborate this claim.
ER -