Non-line-of-sight imaging and tracking of moving objects based on deep learning

Detection of objects outside the line of sight remains a challenge in many practical applications. There have been various researches realizing 2D or 3D imaging of static hidden objects, whose aim are to improve the resolution of reconstructed images. While when it comes to the tracking of continuou...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Optics express 2022-05, Vol.30 (10), p.16758-16772
Hauptverfasser: He, JinHui, Wu, ShuKong, Wei, Ran, Zhang, YuNing
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 16772
container_issue 10
container_start_page 16758
container_title Optics express
container_volume 30
creator He, JinHui
Wu, ShuKong
Wei, Ran
Zhang, YuNing
description Detection of objects outside the line of sight remains a challenge in many practical applications. There have been various researches realizing 2D or 3D imaging of static hidden objects, whose aim are to improve the resolution of reconstructed images. While when it comes to the tracking of continuously moving objects, the speed of imaging and the accuracy of positioning becomes the priorities to optimize. Previous works have achieved centimeter-level or even higher precision of positioning through marking coordinates in intervals of 3 seconds to tens of milliseconds. Here a deep learning framework is proposed to realize the imaging and dynamic tracking of targets simultaneously using a standard RGB camera. Through simulation experiments, we firstly use the designed neural network to achieve positioning of a 3D mannequin with sub-centimeter accuracy (relative error under 1.8%), costing only 3 milliseconds per estimation in average. Furthermore, we apply the system to a physical scene to successfully recover the video signal of the moving target, intuitively revealing its trajectory. We demonstrate an efficient and inexpensive approach that can present the movement of objects around the corner in real time, profiting from the imaging of the NLOS scene, it is also possible to identify the hidden target. This technique can be ultilized to security surveillance, military reconnaissance, autonomous driving and other fields.
doi_str_mv 10.1364/OE.455803
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2724239570</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2724239570</sourcerecordid><originalsourceid>FETCH-LOGICAL-c1423-d75786b3fba5d68b56f22c37e3ff21e316ca893ed4791498bda67de3929025af3</originalsourceid><addsrcrecordid>eNpNUMtOwzAQtBBIlMKBP_ARDi5-xskRVeUhVeQCZ8ux1yUljYudIvH3pIQDp53VjmZmB6FrRhdMFPKuXi2kUiUVJ2jGaCWJpKU-_YfP0UXOW0qZ1JWeofol9qRreyAxkNxu3gfc7uym7TfY9h4PybqP4xID3sWvX9RswQ0ZNzaDx7HHHmCPO7CpH8-X6CzYLsPV35yjt4fV6_KJrOvH5-X9mjgmuSBeK10WjQiNVb4oG1UEzp3QIELgDAQrnC0rAX5MyWRVNt4W2oOoeEW5skHM0c2ku0_x8wB5MLs2O-g620M8ZMM1H30qpelIvZ2oLsWcEwSzT-OP6dswao6lmXplptLEDwgpXf0</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2724239570</pqid></control><display><type>article</type><title>Non-line-of-sight imaging and tracking of moving objects based on deep learning</title><source>DOAJ Directory of Open Access Journals</source><source>EZB-FREE-00999 freely available EZB journals</source><source>Alma/SFX Local Collection</source><creator>He, JinHui ; Wu, ShuKong ; Wei, Ran ; Zhang, YuNing</creator><creatorcontrib>He, JinHui ; Wu, ShuKong ; Wei, Ran ; Zhang, YuNing</creatorcontrib><description>Detection of objects outside the line of sight remains a challenge in many practical applications. There have been various researches realizing 2D or 3D imaging of static hidden objects, whose aim are to improve the resolution of reconstructed images. While when it comes to the tracking of continuously moving objects, the speed of imaging and the accuracy of positioning becomes the priorities to optimize. Previous works have achieved centimeter-level or even higher precision of positioning through marking coordinates in intervals of 3 seconds to tens of milliseconds. Here a deep learning framework is proposed to realize the imaging and dynamic tracking of targets simultaneously using a standard RGB camera. Through simulation experiments, we firstly use the designed neural network to achieve positioning of a 3D mannequin with sub-centimeter accuracy (relative error under 1.8%), costing only 3 milliseconds per estimation in average. Furthermore, we apply the system to a physical scene to successfully recover the video signal of the moving target, intuitively revealing its trajectory. We demonstrate an efficient and inexpensive approach that can present the movement of objects around the corner in real time, profiting from the imaging of the NLOS scene, it is also possible to identify the hidden target. This technique can be ultilized to security surveillance, military reconnaissance, autonomous driving and other fields.</description><identifier>ISSN: 1094-4087</identifier><identifier>EISSN: 1094-4087</identifier><identifier>DOI: 10.1364/OE.455803</identifier><language>eng</language><ispartof>Optics express, 2022-05, Vol.30 (10), p.16758-16772</ispartof><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c1423-d75786b3fba5d68b56f22c37e3ff21e316ca893ed4791498bda67de3929025af3</citedby><cites>FETCH-LOGICAL-c1423-d75786b3fba5d68b56f22c37e3ff21e316ca893ed4791498bda67de3929025af3</cites><orcidid>0000-0001-5855-1129 ; 0000-0003-2213-2859</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,776,780,860,27901,27902</link.rule.ids></links><search><creatorcontrib>He, JinHui</creatorcontrib><creatorcontrib>Wu, ShuKong</creatorcontrib><creatorcontrib>Wei, Ran</creatorcontrib><creatorcontrib>Zhang, YuNing</creatorcontrib><title>Non-line-of-sight imaging and tracking of moving objects based on deep learning</title><title>Optics express</title><description>Detection of objects outside the line of sight remains a challenge in many practical applications. There have been various researches realizing 2D or 3D imaging of static hidden objects, whose aim are to improve the resolution of reconstructed images. While when it comes to the tracking of continuously moving objects, the speed of imaging and the accuracy of positioning becomes the priorities to optimize. Previous works have achieved centimeter-level or even higher precision of positioning through marking coordinates in intervals of 3 seconds to tens of milliseconds. Here a deep learning framework is proposed to realize the imaging and dynamic tracking of targets simultaneously using a standard RGB camera. Through simulation experiments, we firstly use the designed neural network to achieve positioning of a 3D mannequin with sub-centimeter accuracy (relative error under 1.8%), costing only 3 milliseconds per estimation in average. Furthermore, we apply the system to a physical scene to successfully recover the video signal of the moving target, intuitively revealing its trajectory. We demonstrate an efficient and inexpensive approach that can present the movement of objects around the corner in real time, profiting from the imaging of the NLOS scene, it is also possible to identify the hidden target. This technique can be ultilized to security surveillance, military reconnaissance, autonomous driving and other fields.</description><issn>1094-4087</issn><issn>1094-4087</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNpNUMtOwzAQtBBIlMKBP_ARDi5-xskRVeUhVeQCZ8ux1yUljYudIvH3pIQDp53VjmZmB6FrRhdMFPKuXi2kUiUVJ2jGaCWJpKU-_YfP0UXOW0qZ1JWeofol9qRreyAxkNxu3gfc7uym7TfY9h4PybqP4xID3sWvX9RswQ0ZNzaDx7HHHmCPO7CpH8-X6CzYLsPV35yjt4fV6_KJrOvH5-X9mjgmuSBeK10WjQiNVb4oG1UEzp3QIELgDAQrnC0rAX5MyWRVNt4W2oOoeEW5skHM0c2ku0_x8wB5MLs2O-g620M8ZMM1H30qpelIvZ2oLsWcEwSzT-OP6dswao6lmXplptLEDwgpXf0</recordid><startdate>20220509</startdate><enddate>20220509</enddate><creator>He, JinHui</creator><creator>Wu, ShuKong</creator><creator>Wei, Ran</creator><creator>Zhang, YuNing</creator><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0001-5855-1129</orcidid><orcidid>https://orcid.org/0000-0003-2213-2859</orcidid></search><sort><creationdate>20220509</creationdate><title>Non-line-of-sight imaging and tracking of moving objects based on deep learning</title><author>He, JinHui ; Wu, ShuKong ; Wei, Ran ; Zhang, YuNing</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c1423-d75786b3fba5d68b56f22c37e3ff21e316ca893ed4791498bda67de3929025af3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>He, JinHui</creatorcontrib><creatorcontrib>Wu, ShuKong</creatorcontrib><creatorcontrib>Wei, Ran</creatorcontrib><creatorcontrib>Zhang, YuNing</creatorcontrib><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><jtitle>Optics express</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>He, JinHui</au><au>Wu, ShuKong</au><au>Wei, Ran</au><au>Zhang, YuNing</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Non-line-of-sight imaging and tracking of moving objects based on deep learning</atitle><jtitle>Optics express</jtitle><date>2022-05-09</date><risdate>2022</risdate><volume>30</volume><issue>10</issue><spage>16758</spage><epage>16772</epage><pages>16758-16772</pages><issn>1094-4087</issn><eissn>1094-4087</eissn><abstract>Detection of objects outside the line of sight remains a challenge in many practical applications. There have been various researches realizing 2D or 3D imaging of static hidden objects, whose aim are to improve the resolution of reconstructed images. While when it comes to the tracking of continuously moving objects, the speed of imaging and the accuracy of positioning becomes the priorities to optimize. Previous works have achieved centimeter-level or even higher precision of positioning through marking coordinates in intervals of 3 seconds to tens of milliseconds. Here a deep learning framework is proposed to realize the imaging and dynamic tracking of targets simultaneously using a standard RGB camera. Through simulation experiments, we firstly use the designed neural network to achieve positioning of a 3D mannequin with sub-centimeter accuracy (relative error under 1.8%), costing only 3 milliseconds per estimation in average. Furthermore, we apply the system to a physical scene to successfully recover the video signal of the moving target, intuitively revealing its trajectory. We demonstrate an efficient and inexpensive approach that can present the movement of objects around the corner in real time, profiting from the imaging of the NLOS scene, it is also possible to identify the hidden target. This technique can be ultilized to security surveillance, military reconnaissance, autonomous driving and other fields.</abstract><doi>10.1364/OE.455803</doi><tpages>15</tpages><orcidid>https://orcid.org/0000-0001-5855-1129</orcidid><orcidid>https://orcid.org/0000-0003-2213-2859</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 1094-4087
ispartof Optics express, 2022-05, Vol.30 (10), p.16758-16772
issn 1094-4087
1094-4087
language eng
recordid cdi_proquest_miscellaneous_2724239570
source DOAJ Directory of Open Access Journals; EZB-FREE-00999 freely available EZB journals; Alma/SFX Local Collection
title Non-line-of-sight imaging and tracking of moving objects based on deep learning
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-07T22%3A01%3A21IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Non-line-of-sight%20imaging%20and%20tracking%20of%20moving%20objects%20based%20on%20deep%20learning&rft.jtitle=Optics%20express&rft.au=He,%20JinHui&rft.date=2022-05-09&rft.volume=30&rft.issue=10&rft.spage=16758&rft.epage=16772&rft.pages=16758-16772&rft.issn=1094-4087&rft.eissn=1094-4087&rft_id=info:doi/10.1364/OE.455803&rft_dat=%3Cproquest_cross%3E2724239570%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2724239570&rft_id=info:pmid/&rfr_iscdi=true