Fine Tuning Large Models for Straw Detection in Harvested Fields Under Few-Shot Learning Scenarios

The effective monitoring of crop residues, particularly straw in non-harvested fields, is essential for sustainable agricultural practices and environmental management. Traditional methods of straw detection often face challenges due to limited training data, high annotation complexity, and the need...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:IEEE geoscience and remote sensing letters Ročník 22; s. 1 - 5
Hlavní autori: Wu, Di, Liu, Xi, Bai, Song, Liu, Caixia
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Piscataway IEEE 2025
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Predmet:
ISSN:1545-598X, 1558-0571
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract The effective monitoring of crop residues, particularly straw in non-harvested fields, is essential for sustainable agricultural practices and environmental management. Traditional methods of straw detection often face challenges due to limited training data, high annotation complexity, and the need for accurate feature recognition. In response to these challenges, this study investigates the effectiveness of the segment anything model (SAM)-vision transformer (ViT)-huge-low-rank adaptation (LoRA) method, which leverages few-shot learning techniques to accurately and efficiently identify straw using only 0.65% of the available training data. A comparative analysis was performed under consistent testing conditions against several established algorithms, including Deeplabv3, FCN, PSPNet, TransformerUNet, UNet3+, AFFormer, and DynaMas. The results indicate that the SAM-ViT-huge-LoRA method achieves an <inline-formula> <tex-math notation="LaTeX">F1 </tex-math></inline-formula>-score of 83.6%, exceeding the performance of the second best algorithm by at least 2%. Furthermore, the method demonstrates an intersection over union (IoU) metric of 98.02%, surpassing competing models by a minimum of 25%. This research highlights the potential of few-shot learning in scenarios characterized by data scarcity and complex annotation processes. By effectively fine-tuning large models with a small amount of high-quality training data, our approach addresses the challenges of insufficient sample sizes, optimizes the use of limited datasets, reduces annotation costs, and significantly enhances recognition accuracy.
AbstractList The effective monitoring of crop residues, particularly straw in non-harvested fields, is essential for sustainable agricultural practices and environmental management. Traditional methods of straw detection often face challenges due to limited training data, high annotation complexity, and the need for accurate feature recognition. In response to these challenges, this study investigates the effectiveness of the segment anything model (SAM)-vision transformer (ViT)-huge-low-rank adaptation (LoRA) method, which leverages few-shot learning techniques to accurately and efficiently identify straw using only 0.65% of the available training data. A comparative analysis was performed under consistent testing conditions against several established algorithms, including Deeplabv3, FCN, PSPNet, TransformerUNet, UNet3+, AFFormer, and DynaMas. The results indicate that the SAM-ViT-huge-LoRA method achieves an [Formula Omitted]-score of 83.6%, exceeding the performance of the second best algorithm by at least 2%. Furthermore, the method demonstrates an intersection over union (IoU) metric of 98.02%, surpassing competing models by a minimum of 25%. This research highlights the potential of few-shot learning in scenarios characterized by data scarcity and complex annotation processes. By effectively fine-tuning large models with a small amount of high-quality training data, our approach addresses the challenges of insufficient sample sizes, optimizes the use of limited datasets, reduces annotation costs, and significantly enhances recognition accuracy.
The effective monitoring of crop residues, particularly straw in non-harvested fields, is essential for sustainable agricultural practices and environmental management. Traditional methods of straw detection often face challenges due to limited training data, high annotation complexity, and the need for accurate feature recognition. In response to these challenges, this study investigates the effectiveness of the segment anything model (SAM)-vision transformer (ViT)-huge-low-rank adaptation (LoRA) method, which leverages few-shot learning techniques to accurately and efficiently identify straw using only 0.65% of the available training data. A comparative analysis was performed under consistent testing conditions against several established algorithms, including Deeplabv3, FCN, PSPNet, TransformerUNet, UNet3+, AFFormer, and DynaMas. The results indicate that the SAM-ViT-huge-LoRA method achieves an <inline-formula> <tex-math notation="LaTeX">F1 </tex-math></inline-formula>-score of 83.6%, exceeding the performance of the second best algorithm by at least 2%. Furthermore, the method demonstrates an intersection over union (IoU) metric of 98.02%, surpassing competing models by a minimum of 25%. This research highlights the potential of few-shot learning in scenarios characterized by data scarcity and complex annotation processes. By effectively fine-tuning large models with a small amount of high-quality training data, our approach addresses the challenges of insufficient sample sizes, optimizes the use of limited datasets, reduces annotation costs, and significantly enhances recognition accuracy.
Author Liu, Caixia
Wu, Di
Liu, Xi
Bai, Song
Author_xml – sequence: 1
  givenname: Di
  surname: Wu
  fullname: Wu, Di
  organization: School of Surveying and Mapping Engineering, Heilongjiang Institute of Technology, Harbin, China
– sequence: 2
  givenname: Xi
  surname: Liu
  fullname: Liu, Xi
  organization: Heilongjiang Geomatics Centre of Ministry of Natural Resources, Harbin, China
– sequence: 3
  givenname: Song
  surname: Bai
  fullname: Bai, Song
  organization: Heilongjiang Geomatics Centre of Ministry of Natural Resources, Harbin, China
– sequence: 4
  givenname: Caixia
  orcidid: 0000-0001-5851-6374
  surname: Liu
  fullname: Liu, Caixia
  email: liucx@radi.ac.cn
  organization: State Key Laboratory of Remote Sensing Science and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China
BookMark eNpNkE9Lw0AQxRepYKt-AMHDgufU_ZvNHqXaVogItgVvYZPM1pS6W3dTi9_exPbgaWbgvTeP3wgNnHeA0A0lY0qJvs9nb4sxI0yOuRQZJekZGlIps4RIRQf9LmQidfZ-gUYxbghhIsvUEJXTxgFe7l3j1jg3YQ34xdewjdj6gBdtMAf8CC1UbeMdbhyem_ANsYUaTxvY1hGvXA0BT-GQLD58i3Mw4S9sUYEzofHxCp1bs41wfZqXaDV9Wk7mSf46e5485EnVdWkTQ3WpM2tSBpUhDIhVqRDW0FRZxUqrBOeqrtM6M4pIxquyu6zUihEltDD8Et0dc3fBf-27jsXG74PrXhacUZ1pRjnvVPSoqoKPMYAtdqH5NOGnoKToURY9yqJHWZxQdp7bo6cBgH96zaQQiv8CjXVxUg
CODEN IGRSBY
Cites_doi 10.1109/CVPR.2017.549
10.1007/978-3-031-34048-2_58
10.1609/aaai.v37i1.25126
10.1080/26395940.2021.1948354
10.1145/3582688
10.1109/CVPR.2017.660
10.1038/s41598-024-60375-1
10.1109/CVPR.2015.7298965
10.1016/j.media.2024.103280
10.1109/CVPR52729.2023.01085
10.1007/978-3-030-00889-5_1
10.1016/j.rse.2011.09.016
10.3390/rs13071358
10.1007/978-3-319-50835-1_22
10.3390/rs16020342
10.7717/peerj-cs.432
10.3390/su11061762
10.3390/s19183859
ContentType Journal Article
Copyright Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
Copyright_xml – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2025
DBID 97E
RIA
RIE
AAYXX
CITATION
7SC
7SP
7TG
7UA
8FD
C1K
F1W
FR3
H8D
H96
JQ2
KL.
KR7
L.G
L7M
L~C
L~D
DOI 10.1109/LGRS.2025.3548106
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005-present
IEEE All-Society Periodicals Package (ASPP) 1998-Present
IEEE Electronic Library (IEL)
CrossRef
Computer and Information Systems Abstracts
Electronics & Communications Abstracts
Meteorological & Geoastrophysical Abstracts
Water Resources Abstracts
Technology Research Database
Environmental Sciences and Pollution Management
ASFA: Aquatic Sciences and Fisheries Abstracts
Engineering Research Database
Aerospace Database
Aquatic Science & Fisheries Abstracts (ASFA) 2: Ocean Technology, Policy & Non-Living Resources
ProQuest Computer Science Collection
Meteorological & Geoastrophysical Abstracts - Academic
Civil Engineering Abstracts
Aquatic Science & Fisheries Abstracts (ASFA) Professional
Advanced Technologies Database with Aerospace
Computer and Information Systems Abstracts – Academic
Computer and Information Systems Abstracts Professional
DatabaseTitle CrossRef
Civil Engineering Abstracts
Aquatic Science & Fisheries Abstracts (ASFA) Professional
Technology Research Database
Computer and Information Systems Abstracts – Academic
Electronics & Communications Abstracts
ProQuest Computer Science Collection
Computer and Information Systems Abstracts
Water Resources Abstracts
Environmental Sciences and Pollution Management
Computer and Information Systems Abstracts Professional
Aerospace Database
Meteorological & Geoastrophysical Abstracts
Aquatic Science & Fisheries Abstracts (ASFA) 2: Ocean Technology, Policy & Non-Living Resources
ASFA: Aquatic Sciences and Fisheries Abstracts
Engineering Research Database
Advanced Technologies Database with Aerospace
Meteorological & Geoastrophysical Abstracts - Academic
DatabaseTitleList Civil Engineering Abstracts

Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Xplore
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Geography
Geology
EISSN 1558-0571
EndPage 5
ExternalDocumentID 10_1109_LGRS_2025_3548106
10925447
Genre orig-research
GrantInformation_xml – fundername: Basic Scientific Research Operating Expenses of Heilongjiang Provincial Universities of China
  grantid: 2021GJ05
  funderid: 10.13039/501100017600
– fundername: Natural Science Foundation of Heilongjiang Province of China
  grantid: LH2022C076
  funderid: 10.13039/501100005046
GroupedDBID 0R~
29I
4.4
5GY
5VS
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACIWK
AENEX
AETIX
AFRAH
AGQYO
AGSQL
AHBIQ
AIBXA
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
EBS
EJD
HZ~
H~9
IFIPE
IPLJI
JAVBF
LAI
M43
O9-
OCL
P2P
RIA
RIE
RNS
~02
AAYXX
CITATION
7SC
7SP
7TG
7UA
8FD
C1K
F1W
FR3
H8D
H96
JQ2
KL.
KR7
L.G
L7M
L~C
L~D
ID FETCH-LOGICAL-c248t-a19b98fa62eca02e0f7644fa167f72bf74337dd6d8a70523cb7ddf597207494a3
IEDL.DBID RIE
ISICitedReferencesCount 0
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001510911000005&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1545-598X
IngestDate Thu Oct 16 08:36:23 EDT 2025
Sat Nov 29 07:49:23 EST 2025
Wed Aug 27 01:45:44 EDT 2025
IsPeerReviewed true
IsScholarly true
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c248t-a19b98fa62eca02e0f7644fa167f72bf74337dd6d8a70523cb7ddf597207494a3
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ORCID 0000-0001-5851-6374
PQID 3219892133
PQPubID 75725
PageCount 5
ParticipantIDs proquest_journals_3219892133
ieee_primary_10925447
crossref_primary_10_1109_LGRS_2025_3548106
PublicationCentury 2000
PublicationDate 20250000
2025-00-00
20250101
PublicationDateYYYYMMDD 2025-01-01
PublicationDate_xml – year: 2025
  text: 20250000
PublicationDecade 2020
PublicationPlace Piscataway
PublicationPlace_xml – name: Piscataway
PublicationTitle IEEE geoscience and remote sensing letters
PublicationTitleAbbrev LGRS
PublicationYear 2025
Publisher IEEE
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml – name: IEEE
– name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References ref13
ref12
Hu (ref8) 2021
ref14
ref20
ref11
ref10
ref21
ref2
ref1
ref17
ref19
ref18
Loshchilov (ref15) 2017
ref7
ref9
ref4
ref3
Chen (ref16) 2017
ref6
ref5
References_xml – ident: ref13
  doi: 10.1109/CVPR.2017.549
– ident: ref10
  doi: 10.1007/978-3-031-34048-2_58
– ident: ref20
  doi: 10.1609/aaai.v37i1.25126
– ident: ref1
  doi: 10.1080/26395940.2021.1948354
– ident: ref11
  doi: 10.1145/3582688
– ident: ref17
  doi: 10.1109/CVPR.2017.660
– ident: ref9
  doi: 10.1038/s41598-024-60375-1
– ident: ref14
  doi: 10.1109/CVPR.2015.7298965
– ident: ref18
  doi: 10.1016/j.media.2024.103280
– ident: ref21
  doi: 10.1109/CVPR52729.2023.01085
– ident: ref19
  doi: 10.1007/978-3-030-00889-5_1
– year: 2017
  ident: ref15
  article-title: Decoupled weight decay regularization
  publication-title: arXiv:1711.05101
– ident: ref4
  doi: 10.1016/j.rse.2011.09.016
– year: 2017
  ident: ref16
  article-title: Rethinking atrous convolution for semantic image segmentation
  publication-title: arXiv:1706.05587
– ident: ref6
  doi: 10.3390/rs13071358
– ident: ref12
  doi: 10.1007/978-3-319-50835-1_22
– ident: ref2
  doi: 10.3390/rs16020342
– ident: ref5
  doi: 10.7717/peerj-cs.432
– year: 2021
  ident: ref8
  article-title: LoRA: Low-rank adaptation of large language models
  publication-title: arXiv:2106.09685
– ident: ref3
  doi: 10.3390/su11061762
– ident: ref7
  doi: 10.3390/s19183859
SSID ssj0024887
Score 2.4090767
Snippet The effective monitoring of crop residues, particularly straw in non-harvested fields, is essential for sustainable agricultural practices and environmental...
SourceID proquest
crossref
ieee
SourceType Aggregation Database
Index Database
Publisher
StartPage 1
SubjectTerms Adaptation models
Agricultural practices
Algorithms
Annotations
Bibliographic information
Comparative analysis
Complexity
Computational modeling
Crop residues
Crops
Effectiveness
Environmental management
Feature recognition
Few-shot learning
Harvesting
Image segmentation
Learning
low-rank adaptation (LoRA)
Machine learning
Monitoring
Remote sensing
segment anything model (SAM)
semantic segmentation
Straw
Sustainable agriculture
Sustainable practices
Training
Training data
Transformers
Vectors
Title Fine Tuning Large Models for Straw Detection in Harvested Fields Under Few-Shot Learning Scenarios
URI https://ieeexplore.ieee.org/document/10925447
https://www.proquest.com/docview/3219892133
Volume 22
WOSCitedRecordID wos001510911000005&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Xplore
  customDbUrl:
  eissn: 1558-0571
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0024887
  issn: 1545-598X
  databaseCode: RIE
  dateStart: 20040101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NS8QwEB1UFL34La5f5OBJiHbTbZMcRa0elkVclb2VtJu4C9KVbVfx3zuTVlwQD95aSEqZaZo3k5n3AE61FHEknOSZ7gw54n_BzVAZnqsgM4hHnPGUQs9d2eupwUDfN83qvhfGWuuLz-w5Xfqz_OEkn1GqDFe4JkYtuQiLUsZ1s9YPsZ7yangECXik1aA5wsQ5F93bhz6GgiI6DxGgt0ndaG4T8qoqv37Ffn9JNv75Zpuw3gBJdll7fgsWbLENq42m-ehzG1ZuvWjv5w5kCUJJ9jijFAjrUuk3Iwm015IhYmXET_vBrm3li7IKNi4YCQb5PChLqMCtZF4diSX2g_dHk4o1pKwvrJ_bAoPtSbkLT8nN49Udb7QVeI6Wqrhp60wrZ2JhcxMIGziJyMiZdiydFJlDYBFKUptSRlLmOM_wzmH0IRBz6I4J92CpmBR2H5h0JgqlynGY6xg31MrEgVNGuyjKAtVpwdm3sdO3mkIj9aFHoFPyTEqeSRvPtGCXrDs3sDZsC46-_ZM2q6xMQ0EVXwLD7IM_ph3CGj29zpkcwVI1ndljWM7fq3E5PfEf0BdOl8KX
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Rb9MwED6xwVRegHVFFArzw56QvKVOU9uPCEiLFqppLVPfIie110ooRUtK1X_PnZOKSmgPPCWRbCW6s-PvzufvA7jQUgwj4STP9GDBEf8LbhbK8FwFmUE84oynFLpL5GSi5nN90xxW92dhrLW--Mxe0q3fy1-s8w2lynCGa2LUkkfwFC8iqI9r_aXWU14Pj0ABj7SaN5uY2OsqGd1OMRgU0WWIEL1P-kYHy5DXVfnnZ-xXmPjlf37bK3jRQEn2qfb9KTyxRRtajar5cteGk5GX7d2dQRYjmGSzDSVBWELF34xE0H6WDDErI4baLftiK1-WVbBVwUgyyGdCWUwlbiXz-kgstls-Xa4r1tCy3rNpbgsMt9dlB37EX2efx7xRV-A5Wqripq8zrZwZCpubQNjAScRGzvSH0kmROYQWoSS9KWUk5Y7zDJ8cxh8CUYcemPA1HBfrwr4BJp2JQqlybOYGxi20MsPAKaNdFGWBGnTh497Y6a-aRCP1wUegU_JMSp5JG890oUPWPWhYG7YLvb1_0maelWkoqOZLYKD99pFu59Aaz74nafJtcv0OntOb6gxKD46rh419D8_y39WqfPjgB9MfmmTF3g
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=Fine+Tuning+Large+Models+for+Straw+Detection+in+Harvested+Fields+Under+Few-Shot+Learning+Scenarios&rft.jtitle=IEEE+geoscience+and+remote+sensing+letters&rft.au=Wu%2C+Di&rft.au=Liu%2C+Xi&rft.au=Bai%2C+Song&rft.au=Liu%2C+Caixia&rft.date=2025&rft.pub=IEEE&rft.issn=1545-598X&rft.volume=22&rft.spage=1&rft.epage=5&rft_id=info:doi/10.1109%2FLGRS.2025.3548106&rft.externalDocID=10925447
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1545-598X&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1545-598X&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1545-598X&client=summon