Fast Background Subtraction for Through Obstacle Imaging Using a Low Cost, Portable Measurement System

Deniz Kumlu*, Orhan Apaydin, Selçuk Paker, Isin Erer

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Echoes from the obstacle or non-target objects in the imagery scene may suppress the target returns in radar imaging. Thus, a background subtraction procedure is required to enhance the target visibility. Since the conventional subspace or low rank and sparse decomposition methods require singular value decomposition (SVD) computations, in this study we investigate fast background subtraction methods which are more appropriate for field studies conducted by mobile devices. We obtain the low rank component using SVD free procedures using random matrices or reduce the dimension of the raw input prior to SVD calculations. A low-cost portable vector network analyzer (VNA) device and Vivaldi antenna pair are used in laboratory measurements. S21 measurements were used to obtain A-scans which are then concatenated to form B-scan image. The proposed fast methods have been compared to the conventional SVD as well as robust principal component analysis (RPCA). The results validate those similar performances have been achieved compared to RPCA with a % 30-85 decrease in running time.

Original languageEnglish
Title of host publication2022 IEEE Conference on Antenna Measurements and Applications, CAMA 2022
PublisherInstitute of Electrical and Electronics Engineers
ISBN (Electronic)9781665490375
DOIs
Publication statusPublished - 2022
Event2022 IEEE Conference on Antenna Measurements and Applications, CAMA 2022 - Guangzhou, China
Duration: 14 Dec 202217 Dec 2022

Publication series

NameIEEE Conference on Antenna Measurements and Applications, CAMA
Volume2022-December
ISSN (Print)2474-1760
ISSN (Electronic)2643-6795

Conference

Conference2022 IEEE Conference on Antenna Measurements and Applications, CAMA 2022
Country/TerritoryChina
CityGuangzhou
Period14/12/2217/12/22

Bibliographical note

Publisher Copyright:
© 2022 IEEE.

Funding

ACKNOWLEDGMENT This work was funded by the Scientific and Technological Research Council of Turkey (TUBITAK) under Project No.120E234.

FundersFunder number
Türkiye Bilimsel ve Teknolojik Araştırma Kurumu120E234

    Keywords

    • background subtraction
    • non-negative matrix factorization
    • randomized matrix factorizations
    • robust principal component analysis
    • through-obstacle imaging

    Fingerprint

    Dive into the research topics of 'Fast Background Subtraction for Through Obstacle Imaging Using a Low Cost, Portable Measurement System'. Together they form a unique fingerprint.

    Cite this