融合可微分渲染的SAR多视角样本增广

TN95; 合成孔径雷达(SAR)因其全天候、全天时的监测能力在民用和军事领域得到广泛应用.近年来,深度学习已被广泛应用于SAR图像自动解译.然而,由于卫星轨道和观测角度的限制,SAR目标样本面临视角覆盖率不全的问题,这为学习型SAR目标检测识别算法带来了挑战.该文提出一种融合可微分渲染的SAR多视角样本生成方法,结合逆向三维重建和正向渲染技术,通过卷积神经网络(CNN)从少量SAR视角图像中反演目标三维表征,然后利用可微分SAR渲染器(DSR)渲染出更多视角样本,实现样本在角度维的插值.另外,方法的训练过程使用DSR构建目标函数,无需三维真值监督.根据仿真数据的实验结果,该方法能够有效地增加...

Full description

Saved in:
Bibliographic Details
Published in雷达学报 Vol. 13; no. 2; pp. 457 - 470
Main Authors 贾赫成, 蒲欣洋, 王燕妮, 符士磊, 徐丰
Format Journal Article
LanguageChinese
Published 电磁波信息科学教育部重点实验室 复旦大学 上海 200433 2024
Subjects
Multi-view sample generation
多视角样本生成
合成孔径雷达(SAR)
3D reconstruction
Synthetic Aperture Radar(SAR)
可微分SAR渲染器(DSR)
卷积神经网络(CNN)
三维重建
Differentiable SAR Renderer(DSR)
Convolutional Neural Network(CNN)
Online AccessGet full text
ISSN2095-283X
DOI10.12000/JR24011

Cover

Abstract TN95; 合成孔径雷达(SAR)因其全天候、全天时的监测能力在民用和军事领域得到广泛应用.近年来,深度学习已被广泛应用于SAR图像自动解译.然而,由于卫星轨道和观测角度的限制,SAR目标样本面临视角覆盖率不全的问题,这为学习型SAR目标检测识别算法带来了挑战.该文提出一种融合可微分渲染的SAR多视角样本生成方法,结合逆向三维重建和正向渲染技术,通过卷积神经网络(CNN)从少量SAR视角图像中反演目标三维表征,然后利用可微分SAR渲染器(DSR)渲染出更多视角样本,实现样本在角度维的插值.另外,方法的训练过程使用DSR构建目标函数,无需三维真值监督.根据仿真数据的实验结果,该方法能够有效地增加多视角SAR目标图像,并提高小样本条件下典型SAR目标识别率.
AbstractList TN95; 合成孔径雷达(SAR)因其全天候、全天时的监测能力在民用和军事领域得到广泛应用.近年来,深度学习已被广泛应用于SAR图像自动解译.然而,由于卫星轨道和观测角度的限制,SAR目标样本面临视角覆盖率不全的问题,这为学习型SAR目标检测识别算法带来了挑战.该文提出一种融合可微分渲染的SAR多视角样本生成方法,结合逆向三维重建和正向渲染技术,通过卷积神经网络(CNN)从少量SAR视角图像中反演目标三维表征,然后利用可微分SAR渲染器(DSR)渲染出更多视角样本,实现样本在角度维的插值.另外,方法的训练过程使用DSR构建目标函数,无需三维真值监督.根据仿真数据的实验结果,该方法能够有效地增加多视角SAR目标图像,并提高小样本条件下典型SAR目标识别率.
Abstract_FL Synthetic Aperture Radar(SAR)is extensively utilized in civilian and military domains due to its all-weather,all-time monitoring capabilities.In recent years,deep learning has been widely employed to automatically interpret SAR images.However,due to the constraints of satellite orbit and incident angle,SAR target samples face the issue of incomplete view coverage,which poses challenges for learning-based SAR target detection and recognition algorithms.This paper proposes a method for generating multi-view samples of SAR targets by integrating differentiable rendering,combining inverse Three-Dimensional(3D)reconstruction,and forward rendering techniques.By designing a Convolutional Neural Network(CNN),the proposed method inversely infers the 3D representation of targets from limited views of SAR target images and then utilizes a Differentiable SAR Renderer(DSR)to render new samples from more views,achieving sample interpolation in the view dimension.Moreover,the training process of the proposed method constructs the objective function using DSR,eliminating the need for 3D ground-truth supervision.According to experimental results on simulated data,this method can effectively increase the number of multi-view SAR target images and improve the recognition rate of typical SAR targets under few-shot conditions.
Author 贾赫成
王燕妮
徐丰
符士磊
蒲欣洋
AuthorAffiliation 电磁波信息科学教育部重点实验室 复旦大学 上海 200433
AuthorAffiliation_xml – name: 电磁波信息科学教育部重点实验室 复旦大学 上海 200433
Author_FL PU Xinyang
FU Shilei
JIA Hecheng
WANG Yanni
XU Feng
Author_FL_xml – sequence: 1
  fullname: JIA Hecheng
– sequence: 2
  fullname: PU Xinyang
– sequence: 3
  fullname: WANG Yanni
– sequence: 4
  fullname: FU Shilei
– sequence: 5
  fullname: XU Feng
Author_xml – sequence: 1
  fullname: 贾赫成
– sequence: 2
  fullname: 蒲欣洋
– sequence: 3
  fullname: 王燕妮
– sequence: 4
  fullname: 符士磊
– sequence: 5
  fullname: 徐丰
BookMark eNrjYmDJy89LZWAQMDTQMzQyMDDQ9woyMjEwNGRh4DQysDTVNbIwjuBg4C0uzkwyMDYzMjUzMDbhZDB4Ma_36YSOp_3rn-5b97Sj7dmOTc_mT34-qyXYMejpklkvlre9WD7p2YLtz-asebpo3tOd-3kYWNMSc4pTeaE0N0Oom2uIs4euj7-7p7Ojj26xoYGBoW6yoZFpqqGFkVmSebKpYaKhublFWmqKSWqypZm5pUmqWaJxqkWSgWWSibGBRbJlkpFJClDeKMnMODU1xTw50cyYm0EVYm55Yl5aYl56fFZ-aVEe0Mb4nJSKJCMDoN-MgL4zBgD9nFRC
ClassificationCodes TN95
ContentType Journal Article
Copyright Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
Copyright_xml – notice: Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
DBID 2B.
4A8
92I
93N
PSX
TCJ
DOI 10.12000/JR24011
DatabaseName Wanfang Data Journals - Hong Kong
WANFANG Data Centre
Wanfang Data Journals
万方数据期刊 - 香港版
China Online Journals (COJ)
China Online Journals (COJ)
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
DocumentTitle_FL Multi-view Sample Augumentation for SAR Based on Differentiable SAR Renderer
EndPage 470
ExternalDocumentID ldxb202402011
GroupedDBID -0Y
2B.
4A8
5VS
92H
92I
93N
ABJNI
ACGFS
ALMA_UNASSIGNED_HOLDINGS
CCEZO
CUBFJ
GROUPED_DOAJ
IPNFZ
KQ8
PSX
RIG
TCJ
TGT
U1G
U5S
ID FETCH-LOGICAL-s1001-c125e1826b7c51a1778fed4ec96794e6a3e8b09b4308c9b24d8fe2b63eed7ca63
ISSN 2095-283X
IngestDate Thu May 29 04:01:07 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 2
Keywords Multi-view sample generation
多视角样本生成
合成孔径雷达(SAR)
3D reconstruction
Synthetic Aperture Radar(SAR)
可微分SAR渲染器(DSR)
卷积神经网络(CNN)
三维重建
Differentiable SAR Renderer(DSR)
Convolutional Neural Network(CNN)
Language Chinese
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-s1001-c125e1826b7c51a1778fed4ec96794e6a3e8b09b4308c9b24d8fe2b63eed7ca63
PageCount 14
ParticipantIDs wanfang_journals_ldxb202402011
PublicationCentury 2000
PublicationDate 2024
PublicationDateYYYYMMDD 2024-01-01
PublicationDate_xml – year: 2024
  text: 2024
PublicationDecade 2020
PublicationTitle 雷达学报
PublicationTitle_FL Journal of Radars
PublicationYear 2024
Publisher 电磁波信息科学教育部重点实验室 复旦大学 上海 200433
Publisher_xml – name: 电磁波信息科学教育部重点实验室 复旦大学 上海 200433
SSID ssib036256034
ssj0001853618
ssib023646539
ssib036438606
ssib051376368
ssib038075146
ssib058814069
ssib011592811
ssib054421839
Score 2.369639
Snippet TN95; 合成孔径雷达(SAR)因其全天候、全天时的监测能力在民用和军事领域得到广泛应用.近年来,深度学习已被广泛应用于SAR图像自动解译.然而,由于卫星轨道和观测角度的限...
SourceID wanfang
SourceType Aggregation Database
StartPage 457
Title 融合可微分渲染的SAR多视角样本增广
URI https://d.wanfangdata.com.cn/periodical/ldxb202402011
Volume 13
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVAFT
  databaseName: Open Access Digital Library
  issn: 2095-283X
  databaseCode: KQ8
  dateStart: 20120101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: http://grweb.coalliance.org/oadl/oadl.html
  omitProxy: true
  ssIdentifier: ssj0001853618
  providerName: Colorado Alliance of Research Libraries
– providerCode: PRVAON
  databaseName: DOAJ Directory of Open Access Journals
  issn: 2095-283X
  databaseCode: DOA
  dateStart: 20120101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: https://www.doaj.org/
  omitProxy: true
  ssIdentifier: ssj0001853618
  providerName: Directory of Open Access Journals
– providerCode: PRVHPJ
  databaseName: ROAD: Directory of Open Access Scholarly Resources
  issn: 2095-283X
  databaseCode: M~E
  dateStart: 20120101
  customDbUrl:
  isFulltext: true
  dateEnd: 99991231
  titleUrlDefault: https://road.issn.org
  omitProxy: true
  ssIdentifier: ssib054421839
  providerName: ISSN International Centre
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV3NaxQxFA9tvehB_MTP0oM5ldX5yCQvx8x2liLoobbQW9nMzupBtmBXkB48SelFFL0VqSJCL6LgRb3oXzNr_wzfezPbnUXBj8vwJvlN3tdO8pKdvAhxrYCka0yvaGkM3nGC4vstoA9s-h40GAs25xQbt27r5TV1cz1Zn5l93NxdMvTX8-3f7iv5H69iGfqVdsn-g2ePGsUCpNG_eEUP4_WvfCwzkDaTsCSzRNpAAhABHek6RKSZdBmXgAQtMy1TkGlEhO1IG8vMSOskqDtuhWBO0S226QzjmbCMd4FMDT_Ylq7N4IhYExcr06mPA2VmpU0ZDyRDyjC3JB3LAE66ZOxnhiiGIIGolCGoVtCEoBCV3MjbxawJxsDpBIICox1SJlDmhDkiPGtC6GHNNTEzMkSAay58RJMlT7aOIqFqYMh8Y9I8w3K0MpdgOZnbsLHCKVVRDmtJfqjktxLwcWDCsNsMV1l-KmODWunQJe26xKnF2jOExhbNWAVF_MbMFlkiINuSjOgUw7kiq-wfVUcfYZjbwjBvfWpUihtvX9QYYlSV0LuOVlR17MovAyHtwKJDvlcwYqtHs-m04vd7jzxZNaAIcFYci2g9q7Ekgf0xThZsBJP-n84aoEzG43uMfTBenqQXxOoY9ORfezrLIGnkJkpCGs0m09tEKQ7Qj-6B0q9VB95Xa6QYS2peoj8yUZ0_mtS7USvH2-0G_e7gbiMyXD0lTtZTugVXvZ-nxcz2vTPiRCPR51kRHO4_LZ_vls8-lt8-lLs7oy-fRq9f_th7gm9e-W7v8GDn8ODF6M3n0av35dv98uv3c2Ktk622l1v1WSWtLcpi1spxolDQXN2bPAm7oTHQL3qqyK3GEa_Q3bgAH1iv4gBy6yPVw_rI6xhjVJN3dXxezA02B8UFseALDT4MiyCOjYoM9qWaZkL9CCy24XsXxXyt70bdF21tTHnz0p8Al8VxoquVxCtibvjgYXEVY-uhn-cfwE_5M5Ae
linkProvider Directory of Open Access Journals
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=%E8%9E%8D%E5%90%88%E5%8F%AF%E5%BE%AE%E5%88%86%E6%B8%B2%E6%9F%93%E7%9A%84SAR%E5%A4%9A%E8%A7%86%E8%A7%92%E6%A0%B7%E6%9C%AC%E5%A2%9E%E5%B9%BF&rft.jtitle=%E9%9B%B7%E8%BE%BE%E5%AD%A6%E6%8A%A5&rft.au=%E8%B4%BE%E8%B5%AB%E6%88%90&rft.au=%E8%92%B2%E6%AC%A3%E6%B4%8B&rft.au=%E7%8E%8B%E7%87%95%E5%A6%AE&rft.au=%E7%AC%A6%E5%A3%AB%E7%A3%8A&rft.date=2024&rft.pub=%E7%94%B5%E7%A3%81%E6%B3%A2%E4%BF%A1%E6%81%AF%E7%A7%91%E5%AD%A6%E6%95%99%E8%82%B2%E9%83%A8%E9%87%8D%E7%82%B9%E5%AE%9E%E9%AA%8C%E5%AE%A4+%E5%A4%8D%E6%97%A6%E5%A4%A7%E5%AD%A6+%E4%B8%8A%E6%B5%B7+200433&rft.issn=2095-283X&rft.volume=13&rft.issue=2&rft.spage=457&rft.epage=470&rft_id=info:doi/10.12000%2FJR24011&rft.externalDocID=ldxb202402011
thumbnail_s http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=http%3A%2F%2Fwww.wanfangdata.com.cn%2Fimages%2FPeriodicalImages%2Fldxb%2Fldxb.jpg