No Cover Image

Journal article 426 views 81 downloads

Automatic pavement texture recognition using lightweight few-shot learning

Shuo Pan Orcid Logo, Hai Yan Orcid Logo, Zhuo Liu Orcid Logo, Ning Chen, Yinghao Miao, Yue Hou

Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Volume: 381, Issue: 2254

Swansea University Author: Yue Hou

  • 63895.VOR.pdf

    PDF | Version of Record

    © 2023 The Authors. Published by the Royal Society. Distributed under the terms of a Creative Commons Attribution 4.0 Licence (CC BY 4.0).

    Download (1004.02KB)

Check full text

DOI (Published version): 10.1098/rsta.2022.0166

Abstract

Texture is a crucial characteristic of roads, closely related to their performance. The recognition of pavement texture is of great significance for road maintenance professionals to detect potential safety hazards and carry out necessary countermeasures. Although deep learning models have been appl...

Full description

Published in: Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
ISSN: 1364-503X 1471-2962
Published: The Royal Society 2023
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa63895
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2023-07-17T10:04:49Z
last_indexed 2023-07-17T10:04:49Z
id cronfa63895
recordtype SURis
fullrecord <?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>63895</id><entry>2023-07-17</entry><title>Automatic pavement texture recognition using lightweight few-shot learning</title><swanseaauthors><author><sid>92bf566c65343cb3ee04ad963eacf31b</sid><firstname>Yue</firstname><surname>Hou</surname><name>Yue Hou</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-07-17</date><deptcode>CIVL</deptcode><abstract>Texture is a crucial characteristic of roads, closely related to their performance. The recognition of pavement texture is of great significance for road maintenance professionals to detect potential safety hazards and carry out necessary countermeasures. Although deep learning models have been applied for recognition, the scarcity of data has always been a limitation. To address this issue, this paper proposes a few-shot learning model based on the Siamese network for pavement texture recognition with a limited dataset. The model achieved 89.8% accuracy in a four-way five-shot task classifying the pavement textures of dense asphalt concrete, micro surface, open-graded friction course and stone matrix asphalt. To align with engineering practice, global average pooling (GAP) and one-dimensional convolution are implemented, creating lightweight models that save storage and training time. Comparative experiments show that the lightweight model with GAP implemented on dense layers and one-dimensional convolution on convolutional layers reduced storage volume by 94% and training time by 99%, despite a 2.9% decrease in classification accuracy. Moreover, the model with only GAP implemented on dense layers achieved the highest accuracy at 93.5%, while reducing storage volume and training time by 83% and 6%, respectively.</abstract><type>Journal Article</type><journal>Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences</journal><volume>381</volume><journalNumber>2254</journalNumber><paginationStart/><paginationEnd/><publisher>The Royal Society</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>1364-503X</issnPrint><issnElectronic>1471-2962</issnElectronic><keywords>Pavement detection, deep learning, convolutional neural network, Siamese network, one-dimensional convolution</keywords><publishedDay>4</publishedDay><publishedMonth>9</publishedMonth><publishedYear>2023</publishedYear><publishedDate>2023-09-04</publishedDate><doi>10.1098/rsta.2022.0166</doi><url>http://dx.doi.org/10.1098/rsta.2022.0166</url><notes/><college>COLLEGE NANME</college><department>Civil Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>CIVL</DepartmentCode><institution>Swansea University</institution><apcterm>SU Library paid the OA fee (TA Institutional Deal)</apcterm><funders>This work was supported by Key Science and Technology Projects in the Transportation Industry in 2021 (grant no. 2021-ZD2-047), Key Science and Technology Projects in the Transportation Industry in 2022 (grant no. 2022-ZD6-116), Plan Project of Shandong Transportation S&amp;T (grant no. 2021B49), Opening Project Fund of Materials Service Safety Assessment Facilities (grant no. MSAF-2021-005) and National Natural Science Foundation of China (grant no. 51978048).</funders><projectreference/><lastEdited>2023-08-23T12:08:54.1871330</lastEdited><Created>2023-07-17T10:58:55.3240423</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering</level></path><authors><author><firstname>Shuo</firstname><surname>Pan</surname><orcid>0000-0001-9452-9738</orcid><order>1</order></author><author><firstname>Hai</firstname><surname>Yan</surname><orcid>0000-0003-0660-7228</orcid><order>2</order></author><author><firstname>Zhuo</firstname><surname>Liu</surname><orcid>0000-0001-9356-8989</orcid><order>3</order></author><author><firstname>Ning</firstname><surname>Chen</surname><order>4</order></author><author><firstname>Yinghao</firstname><surname>Miao</surname><order>5</order></author><author><firstname>Yue</firstname><surname>Hou</surname><order>6</order></author></authors><documents><document><filename>63895__28129__51df47a03f5948bca9ffe242585fc74d.pdf</filename><originalFilename>63895.VOR.pdf</originalFilename><uploaded>2023-07-17T11:06:07.5880966</uploaded><type>Output</type><contentLength>1028120</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>© 2023 The Authors. Published by the Royal Society. Distributed under the terms of a Creative Commons Attribution 4.0 Licence (CC BY 4.0).</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling v2 63895 2023-07-17 Automatic pavement texture recognition using lightweight few-shot learning 92bf566c65343cb3ee04ad963eacf31b Yue Hou Yue Hou true false 2023-07-17 CIVL Texture is a crucial characteristic of roads, closely related to their performance. The recognition of pavement texture is of great significance for road maintenance professionals to detect potential safety hazards and carry out necessary countermeasures. Although deep learning models have been applied for recognition, the scarcity of data has always been a limitation. To address this issue, this paper proposes a few-shot learning model based on the Siamese network for pavement texture recognition with a limited dataset. The model achieved 89.8% accuracy in a four-way five-shot task classifying the pavement textures of dense asphalt concrete, micro surface, open-graded friction course and stone matrix asphalt. To align with engineering practice, global average pooling (GAP) and one-dimensional convolution are implemented, creating lightweight models that save storage and training time. Comparative experiments show that the lightweight model with GAP implemented on dense layers and one-dimensional convolution on convolutional layers reduced storage volume by 94% and training time by 99%, despite a 2.9% decrease in classification accuracy. Moreover, the model with only GAP implemented on dense layers achieved the highest accuracy at 93.5%, while reducing storage volume and training time by 83% and 6%, respectively. Journal Article Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 381 2254 The Royal Society 1364-503X 1471-2962 Pavement detection, deep learning, convolutional neural network, Siamese network, one-dimensional convolution 4 9 2023 2023-09-04 10.1098/rsta.2022.0166 http://dx.doi.org/10.1098/rsta.2022.0166 COLLEGE NANME Civil Engineering COLLEGE CODE CIVL Swansea University SU Library paid the OA fee (TA Institutional Deal) This work was supported by Key Science and Technology Projects in the Transportation Industry in 2021 (grant no. 2021-ZD2-047), Key Science and Technology Projects in the Transportation Industry in 2022 (grant no. 2022-ZD6-116), Plan Project of Shandong Transportation S&T (grant no. 2021B49), Opening Project Fund of Materials Service Safety Assessment Facilities (grant no. MSAF-2021-005) and National Natural Science Foundation of China (grant no. 51978048). 2023-08-23T12:08:54.1871330 2023-07-17T10:58:55.3240423 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering Shuo Pan 0000-0001-9452-9738 1 Hai Yan 0000-0003-0660-7228 2 Zhuo Liu 0000-0001-9356-8989 3 Ning Chen 4 Yinghao Miao 5 Yue Hou 6 63895__28129__51df47a03f5948bca9ffe242585fc74d.pdf 63895.VOR.pdf 2023-07-17T11:06:07.5880966 Output 1028120 application/pdf Version of Record true © 2023 The Authors. Published by the Royal Society. Distributed under the terms of a Creative Commons Attribution 4.0 Licence (CC BY 4.0). true eng https://creativecommons.org/licenses/by/4.0/
title Automatic pavement texture recognition using lightweight few-shot learning
spellingShingle Automatic pavement texture recognition using lightweight few-shot learning
Yue Hou
title_short Automatic pavement texture recognition using lightweight few-shot learning
title_full Automatic pavement texture recognition using lightweight few-shot learning
title_fullStr Automatic pavement texture recognition using lightweight few-shot learning
title_full_unstemmed Automatic pavement texture recognition using lightweight few-shot learning
title_sort Automatic pavement texture recognition using lightweight few-shot learning
author_id_str_mv 92bf566c65343cb3ee04ad963eacf31b
author_id_fullname_str_mv 92bf566c65343cb3ee04ad963eacf31b_***_Yue Hou
author Yue Hou
author2 Shuo Pan
Hai Yan
Zhuo Liu
Ning Chen
Yinghao Miao
Yue Hou
format Journal article
container_title Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
container_volume 381
container_issue 2254
publishDate 2023
institution Swansea University
issn 1364-503X
1471-2962
doi_str_mv 10.1098/rsta.2022.0166
publisher The Royal Society
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 Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering
url http://dx.doi.org/10.1098/rsta.2022.0166
document_store_str 1
active_str 0
description Texture is a crucial characteristic of roads, closely related to their performance. The recognition of pavement texture is of great significance for road maintenance professionals to detect potential safety hazards and carry out necessary countermeasures. Although deep learning models have been applied for recognition, the scarcity of data has always been a limitation. To address this issue, this paper proposes a few-shot learning model based on the Siamese network for pavement texture recognition with a limited dataset. The model achieved 89.8% accuracy in a four-way five-shot task classifying the pavement textures of dense asphalt concrete, micro surface, open-graded friction course and stone matrix asphalt. To align with engineering practice, global average pooling (GAP) and one-dimensional convolution are implemented, creating lightweight models that save storage and training time. Comparative experiments show that the lightweight model with GAP implemented on dense layers and one-dimensional convolution on convolutional layers reduced storage volume by 94% and training time by 99%, despite a 2.9% decrease in classification accuracy. Moreover, the model with only GAP implemented on dense layers achieved the highest accuracy at 93.5%, while reducing storage volume and training time by 83% and 6%, respectively.
published_date 2023-09-04T12:08:55Z
_version_ 1775017850756399104
score 11.036706

Similar Items