Target detection in GPR data using joint low-rank and sparsity constraints

Abdesselam Bouzerdoum; Fok Hing Chi Tivive; Canicious Abeynayake

doi:10.1117/12.2228345

Target detection in GPR data using joint low-rank and sparsity constraints

Abdesselam Bouzerdoum^*, Fok Hing Chi Tivive, Canicious Abeynayake

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

In ground penetrating radars, background clutter, which comprises the signals backscattered from the rough, uneven ground surface and the background noise, impairs the visualization of buried objects and subsurface inspections. In this paper, a clutter mitigation method is proposed for target detection. The removal of background clutter is formulated as a constrained optimization problem to obtain a low-rank matrix and a sparse matrix. The low-rank matrix captures the ground surface reflections and the background noise, whereas the sparse matrix contains the target reflections. An optimization method based on split-Bregman algorithm is developed to estimate these two matrices from the input GPR data. Evaluated on real radar data, the proposed method achieves promising results in removing the background clutter and enhancing the target signature.

Original language	English
Title of host publication	Compressive Sensing V
Subtitle of host publication	From Diverse Modalities to Big Data Analytics
Editors	Fauzia Ahmad
Publisher	SPIE
ISBN (Electronic)	9781510600980
DOIs	https://doi.org/10.1117/12.2228345
Publication status	Published - 2016
Externally published	Yes
Event	Compressive Sensing V: From Diverse Modalities to Big Data Analytics - Baltimore, United States Duration: 20 Apr 2016 → 21 Apr 2016

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	9857
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Compressive Sensing V: From Diverse Modalities to Big Data Analytics
Country/Territory	United States
City	Baltimore
Period	20/04/16 → 21/04/16

Keywords

Ground penetrating radar
clutter removal
ground surface reflections mitigation
low-rank
sparsity constraints
split-Bregman method

Access to Document

10.1117/12.2228345

Cite this

@inproceedings{e8947846000c440fb2a33ddd3ec13a27,

title = "Target detection in GPR data using joint low-rank and sparsity constraints",

abstract = "In ground penetrating radars, background clutter, which comprises the signals backscattered from the rough, uneven ground surface and the background noise, impairs the visualization of buried objects and subsurface inspections. In this paper, a clutter mitigation method is proposed for target detection. The removal of background clutter is formulated as a constrained optimization problem to obtain a low-rank matrix and a sparse matrix. The low-rank matrix captures the ground surface reflections and the background noise, whereas the sparse matrix contains the target reflections. An optimization method based on split-Bregman algorithm is developed to estimate these two matrices from the input GPR data. Evaluated on real radar data, the proposed method achieves promising results in removing the background clutter and enhancing the target signature.",

keywords = "Ground penetrating radar, clutter removal, ground surface reflections mitigation, low-rank, sparsity constraints, split-Bregman method",

author = "Abdesselam Bouzerdoum and Tivive, {Fok Hing Chi} and Canicious Abeynayake",

note = "Publisher Copyright: {\textcopyright} 2016 SPIE.; Compressive Sensing V: From Diverse Modalities to Big Data Analytics ; Conference date: 20-04-2016 Through 21-04-2016",

year = "2016",

doi = "10.1117/12.2228345",

language = "English",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Fauzia Ahmad",

booktitle = "Compressive Sensing V",

address = "United States",

}

Bouzerdoum, A, Tivive, FHC & Abeynayake, C 2016, Target detection in GPR data using joint low-rank and sparsity constraints. in F Ahmad (ed.), Compressive Sensing V: From Diverse Modalities to Big Data Analytics., 98570A, Proceedings of SPIE - The International Society for Optical Engineering, vol. 9857, SPIE, Compressive Sensing V: From Diverse Modalities to Big Data Analytics, Baltimore, United States, 20/04/16. https://doi.org/10.1117/12.2228345

Target detection in GPR data using joint low-rank and sparsity constraints. / Bouzerdoum, Abdesselam; Tivive, Fok Hing Chi; Abeynayake, Canicious.
Compressive Sensing V: From Diverse Modalities to Big Data Analytics. ed. / Fauzia Ahmad. SPIE, 2016. 98570A (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9857).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Target detection in GPR data using joint low-rank and sparsity constraints

AU - Bouzerdoum, Abdesselam

AU - Tivive, Fok Hing Chi

AU - Abeynayake, Canicious

PY - 2016

Y1 - 2016

N2 - In ground penetrating radars, background clutter, which comprises the signals backscattered from the rough, uneven ground surface and the background noise, impairs the visualization of buried objects and subsurface inspections. In this paper, a clutter mitigation method is proposed for target detection. The removal of background clutter is formulated as a constrained optimization problem to obtain a low-rank matrix and a sparse matrix. The low-rank matrix captures the ground surface reflections and the background noise, whereas the sparse matrix contains the target reflections. An optimization method based on split-Bregman algorithm is developed to estimate these two matrices from the input GPR data. Evaluated on real radar data, the proposed method achieves promising results in removing the background clutter and enhancing the target signature.

AB - In ground penetrating radars, background clutter, which comprises the signals backscattered from the rough, uneven ground surface and the background noise, impairs the visualization of buried objects and subsurface inspections. In this paper, a clutter mitigation method is proposed for target detection. The removal of background clutter is formulated as a constrained optimization problem to obtain a low-rank matrix and a sparse matrix. The low-rank matrix captures the ground surface reflections and the background noise, whereas the sparse matrix contains the target reflections. An optimization method based on split-Bregman algorithm is developed to estimate these two matrices from the input GPR data. Evaluated on real radar data, the proposed method achieves promising results in removing the background clutter and enhancing the target signature.

KW - Ground penetrating radar

KW - clutter removal

KW - ground surface reflections mitigation

KW - low-rank

KW - sparsity constraints

KW - split-Bregman method

UR - http://www.scopus.com/inward/record.url?scp=84978654524&partnerID=8YFLogxK

U2 - 10.1117/12.2228345

DO - 10.1117/12.2228345

M3 - Conference contribution

AN - SCOPUS:84978654524

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Compressive Sensing V

A2 - Ahmad, Fauzia

PB - SPIE

T2 - Compressive Sensing V: From Diverse Modalities to Big Data Analytics

Y2 - 20 April 2016 through 21 April 2016

ER -

Target detection in GPR data using joint low-rank and sparsity constraints

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this