No Cover Image

Journal article 74 views 3 downloads

Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool

Tulsi Patel, Mark Jones Orcid Logo, Thomas Redfern

Annals of GIS, Pages: 1 - 19

Swansea University Authors: Tulsi Patel, Mark Jones Orcid Logo

  • Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool.pdf

    PDF | Version of Record

    © 2026 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution License.

    Download (7.99MB)

Check full text

DOI (Published version): 10.1080/19475683.2026.2626395

Abstract

Machine learning for remote sensing imaging relies on up-to-date and accurate labels for model training and testing. Labelling remote sensing imagery is time and cost intensive, requiring expert analysis. Previous labelling tools rely on pre-labelled data for training in order to label new unseen da...

Full description

Published in: Annals of GIS
ISSN: 1947-5683 1947-5691
Published: Informa UK Limited 2026
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa71320
first_indexed 2026-01-27T15:31:40Z
last_indexed 2026-02-07T05:29:01Z
id cronfa71320
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2026-02-06T15:19:54.9681612</datestamp><bib-version>v2</bib-version><id>71320</id><entry>2026-01-27</entry><title>Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool</title><swanseaauthors><author><sid>b7d4b35b18e3f66ce9aa77fb9280aa2f</sid><firstname>Tulsi</firstname><surname>Patel</surname><name>Tulsi Patel</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>2e1030b6e14fc9debd5d5ae7cc335562</sid><ORCID>0000-0001-8991-1190</ORCID><firstname>Mark</firstname><surname>Jones</surname><name>Mark Jones</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2026-01-27</date><deptcode>MACS</deptcode><abstract>Machine learning for remote sensing imaging relies on up-to-date and accurate labels for model training and testing. Labelling remote sensing imagery is time and cost intensive, requiring expert analysis. Previous labelling tools rely on pre-labelled data for training in order to label new unseen data. In this work, we define an unsupervised pipeline for finding and labelling geographical areas of similar context and content within Sentinel-2 satellite imagery. Our approach removes limitations of previous methods by utilising segmentation with convolutional and graph neural networks to encode a more robust feature space for image comparison. Unlike previous approaches we segment the image into homogeneous regions of pixels that are grouped based on colour and spatial similarity. Graph neural networks are used to aggregate information about the surrounding segments enabling the feature representation to encode the local neighbourhood whilst preserving its own local information. This reduces outliers in the labelling tool, allows users to label at a granular level, and allows a rotationally invariant semantic relationship at the image level to be formed within the encoding space. Our pipeline achieves high contextual consistency, with similarity scores of SSIM = 0.96 and SAM = 0.21 under context-aware evaluation, demonstrating robust organisation of the feature space for interactive labelling.</abstract><type>Journal Article</type><journal>Annals of GIS</journal><volume>0</volume><journalNumber/><paginationStart>1</paginationStart><paginationEnd>19</paginationEnd><publisher>Informa UK Limited</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>1947-5683</issnPrint><issnElectronic>1947-5691</issnElectronic><keywords>Satellite imagery; remote sensing; graph neural networks; unsupervised clustering; interactive, labelling tool</keywords><publishedDay>3</publishedDay><publishedMonth>2</publishedMonth><publishedYear>2026</publishedYear><publishedDate>2026-02-03</publishedDate><doi>10.1080/19475683.2026.2626395</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm>SU Library paid the OA fee (TA Institutional Deal)</apcterm><funders>This research was supported by EPSRC grant number [EP/S021892/1] and UKHO.</funders><projectreference/><lastEdited>2026-02-06T15:19:54.9681612</lastEdited><Created>2026-01-27T15:26:58.5648802</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Tulsi</firstname><surname>Patel</surname><order>1</order></author><author><firstname>Mark</firstname><surname>Jones</surname><orcid>0000-0001-8991-1190</orcid><order>2</order></author><author><firstname>Thomas</firstname><surname>Redfern</surname><order>3</order></author></authors><documents><document><filename>71320__36187__0f098f90d9784af491d99ec3d9d10c20.pdf</filename><originalFilename>Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool.pdf</originalFilename><uploaded>2026-02-04T09:49:51.6401850</uploaded><type>Output</type><contentLength>8383042</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>&#xA9; 2026 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution License.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>http://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling 2026-02-06T15:19:54.9681612 v2 71320 2026-01-27 Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool b7d4b35b18e3f66ce9aa77fb9280aa2f Tulsi Patel Tulsi Patel true false 2e1030b6e14fc9debd5d5ae7cc335562 0000-0001-8991-1190 Mark Jones Mark Jones true false 2026-01-27 MACS Machine learning for remote sensing imaging relies on up-to-date and accurate labels for model training and testing. Labelling remote sensing imagery is time and cost intensive, requiring expert analysis. Previous labelling tools rely on pre-labelled data for training in order to label new unseen data. In this work, we define an unsupervised pipeline for finding and labelling geographical areas of similar context and content within Sentinel-2 satellite imagery. Our approach removes limitations of previous methods by utilising segmentation with convolutional and graph neural networks to encode a more robust feature space for image comparison. Unlike previous approaches we segment the image into homogeneous regions of pixels that are grouped based on colour and spatial similarity. Graph neural networks are used to aggregate information about the surrounding segments enabling the feature representation to encode the local neighbourhood whilst preserving its own local information. This reduces outliers in the labelling tool, allows users to label at a granular level, and allows a rotationally invariant semantic relationship at the image level to be formed within the encoding space. Our pipeline achieves high contextual consistency, with similarity scores of SSIM = 0.96 and SAM = 0.21 under context-aware evaluation, demonstrating robust organisation of the feature space for interactive labelling. Journal Article Annals of GIS 0 1 19 Informa UK Limited 1947-5683 1947-5691 Satellite imagery; remote sensing; graph neural networks; unsupervised clustering; interactive, labelling tool 3 2 2026 2026-02-03 10.1080/19475683.2026.2626395 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University SU Library paid the OA fee (TA Institutional Deal) This research was supported by EPSRC grant number [EP/S021892/1] and UKHO. 2026-02-06T15:19:54.9681612 2026-01-27T15:26:58.5648802 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Tulsi Patel 1 Mark Jones 0000-0001-8991-1190 2 Thomas Redfern 3 71320__36187__0f098f90d9784af491d99ec3d9d10c20.pdf Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool.pdf 2026-02-04T09:49:51.6401850 Output 8383042 application/pdf Version of Record true © 2026 The Author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution License. true eng http://creativecommons.org/licenses/by/4.0/
title Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
spellingShingle Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
Tulsi Patel
Mark Jones
title_short Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
title_full Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
title_fullStr Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
title_full_unstemmed Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
title_sort Leveraging convolutional and graph networks for an unsupervised remote sensing labelling tool
author_id_str_mv b7d4b35b18e3f66ce9aa77fb9280aa2f
2e1030b6e14fc9debd5d5ae7cc335562
author_id_fullname_str_mv b7d4b35b18e3f66ce9aa77fb9280aa2f_***_Tulsi Patel
2e1030b6e14fc9debd5d5ae7cc335562_***_Mark Jones
author Tulsi Patel
Mark Jones
author2 Tulsi Patel
Mark Jones
Thomas Redfern
format Journal article
container_title Annals of GIS
container_volume 0
container_start_page 1
publishDate 2026
institution Swansea University
issn 1947-5683
1947-5691
doi_str_mv 10.1080/19475683.2026.2626395
publisher Informa UK Limited
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description Machine learning for remote sensing imaging relies on up-to-date and accurate labels for model training and testing. Labelling remote sensing imagery is time and cost intensive, requiring expert analysis. Previous labelling tools rely on pre-labelled data for training in order to label new unseen data. In this work, we define an unsupervised pipeline for finding and labelling geographical areas of similar context and content within Sentinel-2 satellite imagery. Our approach removes limitations of previous methods by utilising segmentation with convolutional and graph neural networks to encode a more robust feature space for image comparison. Unlike previous approaches we segment the image into homogeneous regions of pixels that are grouped based on colour and spatial similarity. Graph neural networks are used to aggregate information about the surrounding segments enabling the feature representation to encode the local neighbourhood whilst preserving its own local information. This reduces outliers in the labelling tool, allows users to label at a granular level, and allows a rotationally invariant semantic relationship at the image level to be formed within the encoding space. Our pipeline achieves high contextual consistency, with similarity scores of SSIM = 0.96 and SAM = 0.21 under context-aware evaluation, demonstrating robust organisation of the feature space for interactive labelling.
published_date 2026-02-03T05:39:30Z
_version_ 1857893350369132544
score 11.097017

Similar Items