融合半监督学习的无监督遥感影像场景分类
自监督学习可以不依赖样本标签对遥感影像进行特征提取, 但是特征分类仍然依赖有监督方法。为了克服有监督特征分类过程的不足, 实现遥感影像特征的无监督自动分类, 本文提出一种融合半监督学习的无监督语义聚类方法。首先, 使用自监督学习提取遥感影像特征, 抽象出图像包含的高层语义信息; 然后, 基于特征相似度寻找每个样本最相似的近邻, 使用在线聚类将相似样本聚为一类, 训练一个线性分类器; 最后, 根据聚类结果为高置信度样本生成伪标签, 构造标注样本集, 使用半监督方法对模型微调。在4个公开遥感影像场景分类数据集EuroSAT、GID、AID和NWPU-RESISC45上进行验证, 分类精度分别达到了...
Saved in:
| Published in | Ce hui xue bao Vol. 51; no. 5; pp. 691 - 702 |
|---|---|
| Main Authors | , , , , |
| Format | Journal Article |
| Language | Chinese English |
| Published |
Beijing
Surveying and Mapping Press
01.05.2022
火箭军工程大学作战保障学院,陕西 西安 710025%61068 部队,陕西 西安 710100 |
| Subjects | |
| Online Access | Get full text |
| ISSN | 1001-1595 1001-1595 |
| DOI | 10.11947/j.AGCS.2022.20210270 |
Cover
| Abstract | 自监督学习可以不依赖样本标签对遥感影像进行特征提取, 但是特征分类仍然依赖有监督方法。为了克服有监督特征分类过程的不足, 实现遥感影像特征的无监督自动分类, 本文提出一种融合半监督学习的无监督语义聚类方法。首先, 使用自监督学习提取遥感影像特征, 抽象出图像包含的高层语义信息; 然后, 基于特征相似度寻找每个样本最相似的近邻, 使用在线聚类将相似样本聚为一类, 训练一个线性分类器; 最后, 根据聚类结果为高置信度样本生成伪标签, 构造标注样本集, 使用半监督方法对模型微调。在4个公开遥感影像场景分类数据集EuroSAT、GID、AID和NWPU-RESISC45上进行验证, 分类精度分别达到了94.84%、63.55%、76.42%和86.24%。本文方法结合了在线聚类和半监督学习的优点, 缓解了已有方法存在的误差积累和样本利用不充分的问题, 在完全不使用标注样本的情况下, 充分利用自监督特征训练分类模型, 对遥感影像进行场景分类, 达到接近有监督学习的分类效果, 具有良好的应用价值。 |
|---|---|
| AbstractList | P237; 自监督学习可以不依赖样本标签对遥感影像进行特征提取,但是特征分类仍然依赖有监督方法.为了克服有监督特征分类过程的不足,实现遥感影像特征的无监督自动分类,本文提出一种融合半监督学习的无监督语义聚类方法.首先,使用自监督学习提取遥感影像特征,抽象出图像包含的高层语义信息;然后,基于特征相似度寻找每个样本最相似的近邻,使用在线聚类将相似样本聚为一类,训练一个线性分类器;最后,根据聚类结果为高置信度样本生成伪标签,构造标注样本集,使用半监督方法对模型微调.在4个公开遥感影像场景分类数据集EuroSAT、GID、AID和NWPU-RESISC45上进行验证,分类精度分别达到了94.84％、63.55％、76.42％和86.24％.本文方法结合了在线聚类和半监督学习的优点,缓解了已有方法存在的误差积累和样本利用不充分的问题,在完全不使用标注样本的情况下,充分利用自监督特征训练分类模型,对遥感影像进行场景分类,达到接近有监督学习的分类效果,具有良好的应用价值. 自监督学习可以不依赖样本标签对遥感影像进行特征提取, 但是特征分类仍然依赖有监督方法。为了克服有监督特征分类过程的不足, 实现遥感影像特征的无监督自动分类, 本文提出一种融合半监督学习的无监督语义聚类方法。首先, 使用自监督学习提取遥感影像特征, 抽象出图像包含的高层语义信息; 然后, 基于特征相似度寻找每个样本最相似的近邻, 使用在线聚类将相似样本聚为一类, 训练一个线性分类器; 最后, 根据聚类结果为高置信度样本生成伪标签, 构造标注样本集, 使用半监督方法对模型微调。在4个公开遥感影像场景分类数据集EuroSAT、GID、AID和NWPU-RESISC45上进行验证, 分类精度分别达到了94.84%、63.55%、76.42%和86.24%。本文方法结合了在线聚类和半监督学习的优点, 缓解了已有方法存在的误差积累和样本利用不充分的问题, 在完全不使用标注样本的情况下, 充分利用自监督特征训练分类模型, 对遥感影像进行场景分类, 达到接近有监督学习的分类效果, 具有良好的应用价值。 |
| Author | 陈雪冰 白坤 慕晓冬 尤轩昂 朱永清 |
| AuthorAffiliation | 火箭军工程大学作战保障学院,陕西 西安 710025%61068 部队,陕西 西安 710100 |
| AuthorAffiliation_xml | – name: 火箭军工程大学作战保障学院,陕西 西安 710025%61068 部队,陕西 西安 710100 |
| Author_FL | ZHU Yongqing CHEN Xuebing MU Xiaodong BAI Kun YOU Xuanang |
| Author_FL_xml | – sequence: 1 fullname: BAI Kun – sequence: 2 fullname: MU Xiaodong – sequence: 3 fullname: CHEN Xuebing – sequence: 4 fullname: ZHU Yongqing – sequence: 5 fullname: YOU Xuanang |
| Author_xml | – sequence: 1 fullname: 白坤 – sequence: 2 fullname: 慕晓冬 – sequence: 3 fullname: 陈雪冰 – sequence: 4 fullname: 朱永清 – sequence: 5 fullname: 尤轩昂 |
| BookMark | eNpNj81Kw0AUhQepYK19BFfiMvXOTGYysyxBq1BwYfdlkkzUIkltLLoVixShqCvxvyC6E-lSxbdxEn0LWyLo5pzL4eNcziwqRHGkEZrHUMFY2s5Sq1KtuRsVAoRMBANxYAoVMQC2MJOs8O-eQeUk2fYAmE0dRmURya-7gTnrm8FJdn2e3T6Y56fP12F21UsvhnnyffiY9u7Nx8gcnZqbt_TyxfSPs9H7HJoO1U6iy79eQo2V5Ya7atXXa2tutW61mRBW4MuAONrzKBU2UTYojrWgXPsioAFhIRM64JSTkGvq2NzXmGPCFZWeDJXgtIQW89p9FYUq2my24m4nGj9s-lsH3mQ1MAAx5hZyrt2Jd7s62fsDCRfAOKO2pD95RWsx |
| ClassificationCodes | P237 |
| ContentType | Journal Article |
| Copyright | May 2022. This work is published under https://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. Copyright © Wanfang Data Co. Ltd. All Rights Reserved. |
| Copyright_xml | – notice: May 2022. This work is published under https://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. – notice: Copyright © Wanfang Data Co. Ltd. All Rights Reserved. |
| DBID | 3V. 7X2 8FE 8FH 8FK ABUWG AEUYN AFKRA ATCPS AZQEC BBNVY BENPR BHPHI BKSAR CCPQU DWQXO GNUQQ HCIFZ LK8 M0K M7P PATMY PCBAR PHGZM PHGZT PIMPY PKEHL PQEST PQGLB PQQKQ PQUKI PRINS PYCSY 2B. 4A8 92I 93N PSX TCJ |
| DOI | 10.11947/j.AGCS.2022.20210270 |
| DatabaseName | ProQuest Central (Corporate) Agricultural Science Collection ProQuest SciTech Collection ProQuest Natural Science Journals ProQuest Central (Alumni) (purchase pre-March 2016) ProQuest Central (Alumni) ProQuest One Sustainability ProQuest Central UK/Ireland Agricultural & Environmental Science Collection ProQuest Central Essentials Biological Science Database ProQuest Central Natural Science Collection Earth, Atmospheric & Aquatic Science Collection ProQuest One Community College ProQuest Central Korea ProQuest Central Student SciTech Premium Collection Biological Sciences Agricultural Science Database Biological Science Database (ProQuest) Environmental Science Database (ProQuest) Earth, Atmospheric & Aquatic Science Database ProQuest Central Premium ProQuest One Academic Publicly Available Content Database ProQuest One Academic Middle East (New) ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Applied & Life Sciences ProQuest One Academic ProQuest One Academic UKI Edition ProQuest Central China Environmental Science Collection 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) |
| DatabaseTitle | Agricultural Science Database Publicly Available Content Database ProQuest Central Student ProQuest One Academic Middle East (New) ProQuest Central Essentials ProQuest Central (Alumni Edition) SciTech Premium Collection ProQuest One Community College ProQuest Natural Science Collection ProQuest Central China Earth, Atmospheric & Aquatic Science Collection ProQuest Central ProQuest One Applied & Life Sciences ProQuest One Sustainability Natural Science Collection ProQuest Central Korea Agricultural & Environmental Science Collection Biological Science Collection ProQuest Central (New) ProQuest Biological Science Collection ProQuest One Academic Eastern Edition Earth, Atmospheric & Aquatic Science Database Agricultural Science Collection Biological Science Database ProQuest SciTech Collection Environmental Science Collection ProQuest One Academic UKI Edition Environmental Science Database ProQuest One Academic ProQuest One Academic (New) ProQuest Central (Alumni) |
| DatabaseTitleList | Agricultural Science Database |
| Database_xml | – sequence: 1 dbid: BENPR name: ProQuest Central url: http://www.proquest.com/pqcentral?accountid=15518 sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Astronomy & Astrophysics |
| DocumentTitle_FL | Unsupervised remote sensing image scene classification based on semi-supervised learning |
| EISSN | 1001-1595 |
| EndPage | 702 |
| ExternalDocumentID | chxb202205008 |
| GrantInformation_xml | – fundername: 国家自然科学基金 funderid: (61773389) |
| GroupedDBID | -01 3V. 5VS 5XA 5XB 7X2 8FE 8FH 8FK 92E 92I ABJNI ABUWG ACGFS AEUYN AFKRA ALMA_UNASSIGNED_HOLDINGS ATCPS AZQEC BBNVY BENPR BHPHI BKSAR CCEZO CCPQU CCVFK CW9 DWQXO GNUQQ HCIFZ LK8 M0K M7P OK1 P2P PATMY PCBAR PHGZM PHGZT PIMPY PKEHL PQEST PQGLB PQQKQ PQUKI PRINS PYCSY TCJ TGP U1G U5K 2B. 4A8 93N IPNFZ PMFND PSX RIG |
| ID | FETCH-LOGICAL-p588-dc9d27ebb33842a40a61e836ec8d3d25f58ed6362f6e3746ce16126a39b9fa863 |
| IEDL.DBID | BENPR |
| ISSN | 1001-1595 |
| IngestDate | Thu May 29 04:11:08 EDT 2025 Sun Jul 13 04:15:04 EDT 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | true |
| Issue | 5 |
| Keywords | 在线聚类 自监督 遥感影像 场景分类 半监督 |
| Language | Chinese English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-p588-dc9d27ebb33842a40a61e836ec8d3d25f58ed6362f6e3746ce16126a39b9fa863 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| OpenAccessLink | https://www.proquest.com/docview/2680565349?pq-origsite=%requestingapplication%&accountid=15518 |
| PQID | 2680565349 |
| PQPubID | 5229093 |
| PageCount | 12 |
| ParticipantIDs | wanfang_journals_chxb202205008 proquest_journals_2680565349 |
| PublicationCentury | 2000 |
| PublicationDate | 20220501 |
| PublicationDateYYYYMMDD | 2022-05-01 |
| PublicationDate_xml | – month: 05 year: 2022 text: 20220501 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | Beijing |
| PublicationPlace_xml | – name: Beijing |
| PublicationTitle | Ce hui xue bao |
| PublicationTitle_FL | Acta Geodaetica et Cartographica Sinica |
| PublicationYear | 2022 |
| Publisher | Surveying and Mapping Press 火箭军工程大学作战保障学院,陕西 西安 710025%61068 部队,陕西 西安 710100 |
| Publisher_xml | – name: Surveying and Mapping Press – name: 火箭军工程大学作战保障学院,陕西 西安 710025%61068 部队,陕西 西安 710100 |
| SSID | ssib005437539 ssib038074662 ssib051373695 ssib002263888 ssib000862384 ssj0058465 |
| Score | 2.3529463 |
| Snippet | 自监督学习可以不依赖样本标签对遥感影像进行特征提取, 但是特征分类仍然依赖有监督方法。为了克服有监督特征分类过程的不足, 实现遥感影像特征的无监督自动分类, 本文提出... P237; 自监督学习可以不依赖样本标签对遥感影像进行特征提取,但是特征分类仍然依赖有监督方法.为了克服有监督特征分类过程的不足,实现遥感影像特征的无监督自动分类,本文... |
| SourceID | wanfang proquest |
| SourceType | Aggregation Database |
| StartPage | 691 |
| SubjectTerms | Classification Clustering Feature extraction Image classification Labels Learning Remote sensing Semantics Semi-supervised learning |
| Title | 融合半监督学习的无监督遥感影像场景分类 |
| URI | https://www.proquest.com/docview/2680565349 https://d.wanfangdata.com.cn/periodical/chxb202205008 |
| Volume | 51 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVAFT databaseName: Open Access Digital Library customDbUrl: eissn: 1001-1595 dateEnd: 99991231 omitProxy: true ssIdentifier: ssib005437539 issn: 1001-1595 databaseCode: KQ8 dateStart: 20120101 isFulltext: true titleUrlDefault: http://grweb.coalliance.org/oadl/oadl.html providerName: Colorado Alliance of Research Libraries – providerCode: PRVPQU databaseName: ProQuest Central customDbUrl: http://www.proquest.com/pqcentral?accountid=15518 eissn: 1001-1595 dateEnd: 20221231 omitProxy: true ssIdentifier: ssj0058465 issn: 1001-1595 databaseCode: BENPR dateStart: 20100201 isFulltext: true titleUrlDefault: https://www.proquest.com/central providerName: ProQuest |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1LaxRBEC6SDYIXiS-SGJc9iLcxM_2o6T6IrCExKC4-IuQW-jUGIbvRjai5iUGCENST-A5IchPJUcV_4-zqv7B7dja7uXgbZnou1dX9fVXd9RXAOaERbZwFMTzNImZlHCkT6yiTHu0pRxMXCf3rDVy4w64u8aURaPRrYcK1yv6eWGzUtmVCjnyGoPBYzSmTl9buR6FrVDhd7bfQUGVrBXuxkBgbhTESlLEqMHZ5rnHj1iECT8UwoHn_GxLs4ox6_n4A8EGNneFAUI8nNKUoDwh2QG9enKf6oNwTA17WCCWSpTP3LtSvzN72ISgJBV8htErjQzz2yCPVzFTz7hCgzY_DsZKJ1uo91zkOIxsrJ2Ci3g658dbqk9r5WvHcS320T4L882k7f7WVb7_ovn_d_fgl_7r3-_tO991m581O783fp7udzc_5r_382cv8w4_O22_51vPu_s9TsDg_tzi7EJWNF6I17heONdKS1Gntw1dGFIsVJk5QdEZYagnPuHAWPfJl6Ki3j3GeNhJUVGqZKYH0NFSaraabgBpPtDSILpbGUwXMpHAiVok0KaEoFJuE6b49lsvF014eTPUkVEsbDb6alceaFPXBfjan_v__GTgahvYuKE5DZf3BQ3fWk4h1XS09owqj126Kf7ipzww |
| linkProvider | ProQuest |
| linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3NaxQxFA-1RfQiftLWWueg3sZm8vEmORRZa-vWtovoCr0tmSRjEbpb3UqtN7FIEYp6Er8LojeRHlX63zi7-l-YzM52txdvvQ0zZBjyfpP3-73kvYfQBZEAGJz6YngJC5mROFQaJ2EqnbenHDTOA_oLFSjfYTcW-eIA2u3mwvhjld01MV-oTUP7GPkEAeF8NadMXlm5H_quUX53tdtCQxWtFcxkXmKsSOyYs-trTsI1J2evOXtfJGRmujpVDosuA-EKdygxWhoS2yRxWo0RxbCCyAoKVgtDDeEpF9aAW-ZTsDRmoK3jSAQUlYlMlQDqXnsIDTH3YU77DV2drty8tU8vUNHvPx3c--qDcUadXNjjE774O4Ne_T4e0ZiC3OPznizwfPsWR6HjIbxISYokiyfuXS5dn7rtFC_x-WVeycV4H20-vKbqqarf7fOfM8fRsYL4BqUOUk-ggcdLJ9FwqelD8Y3l9eBSkF93Ii3NU0j--biVvdzMtp63371qf_icffv6-8d2--1G6_V2587fJ19aG5-y3Z3s6Yvs_c_Wm-_Z5rP2zq_TqHoQFjiDBuuNuh1GAY8SqQEsltoxE0ilsAKrSOqYUBCKjaCx7nzUin-1WeshawSNF3PUe6qXHiUkT0d21hz9__jz6Ei5ujBfm5-tzJ1FR_2wztnIMTS4-uChPef4y2oyXqAkQLUDxuU_kQsKSA |
| 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%8D%8A%E7%9B%91%E7%9D%A3%E5%AD%A6%E4%B9%A0%E7%9A%84%E6%97%A0%E7%9B%91%E7%9D%A3%E9%81%A5%E6%84%9F%E5%BD%B1%E5%83%8F%E5%9C%BA%E6%99%AF%E5%88%86%E7%B1%BB&rft.jtitle=%E6%B5%8B%E7%BB%98%E5%AD%A6%E6%8A%A5&rft.au=%E7%99%BD%E5%9D%A4&rft.au=%E6%85%95%E6%99%93%E5%86%AC&rft.au=%E9%99%88%E9%9B%AA%E5%86%B0&rft.au=%E6%9C%B1%E6%B0%B8%E6%B8%85&rft.date=2022-05-01&rft.pub=%E7%81%AB%E7%AE%AD%E5%86%9B%E5%B7%A5%E7%A8%8B%E5%A4%A7%E5%AD%A6%E4%BD%9C%E6%88%98%E4%BF%9D%E9%9A%9C%E5%AD%A6%E9%99%A2%2C%E9%99%95%E8%A5%BF+%E8%A5%BF%E5%AE%89+710025%2561068+%E9%83%A8%E9%98%9F%2C%E9%99%95%E8%A5%BF+%E8%A5%BF%E5%AE%89+710100&rft.issn=1001-1595&rft.volume=51&rft.issue=5&rft.spage=691&rft.epage=702&rft_id=info:doi/10.11947%2Fj.AGCS.2022.20210270&rft.externalDocID=chxb202205008 |
| thumbnail_s | http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=http%3A%2F%2Fwww.wanfangdata.com.cn%2Fimages%2FPeriodicalImages%2Fchxb%2Fchxb.jpg |