Deep learning-based concrete defects classification and detection using semantic segmentation
Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity yet increase the accuracy of the damage diagnoses and accessibility in a structural health monitoring (SHM) system. However, despite remarkable...
Gespeichert in:
| Veröffentlicht in: | Structural health monitoring Jg. 23; H. 1; S. 383 |
|---|---|
| Hauptverfasser: | , , |
| Format: | Journal Article |
| Sprache: | Englisch |
| Veröffentlicht: |
England
01.01.2024
|
| Schlagworte: | |
| ISSN: | 1475-9217 |
| Online-Zugang: | Weitere Angaben |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Abstract | Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity yet increase the accuracy of the damage diagnoses and accessibility in a structural health monitoring (SHM) system. However, despite remarkable advances with DL-based SHM, the most significant challenges to achieving the real-time implication are the limited available defects image databases and the selection of DL networks depth. To address these challenges, this research has created a diverse dataset with concrete crack (4087) and spalling (1100) images and used it for damage condition assessment by applying convolutional neural network (CNN) algorithms. CNN-classifier models are used to identify different types of defects and semantic segmentation for labeling the defect patterns within an image. Three CNN-based models-Visual Geometry Group (VGG)19, ResNet50, and InceptionV3 are incorporated as CNN-classifiers. For semantic segmentation, two encoder-decoder models, U-Net and pyramid scene parsing network architecture are developed based on four backbone models, including VGG19, ResNet50, InceptionV3, and EfficientNetB3. The CNN-classifier models are analyzed on two optimizers-stochastic gradient descent (SGD), root mean square propagation (RMSprop), and learning rates-0.1, 0.001, and 0.0001. However, the CNN-segmentation models are analyzed for SGD and adaptive moment estimation, trained with three different learning rates-0.1, 0.01, and 0.0001, and evaluated based on accuracy, intersection over union, precision, recall, and F1-score. InceptionV3 achieves the best performance for defects classification with an accuracy of 91.98% using the RMSprop optimizer. For crack segmentation, EfficientNetB3-based U-Net, and for spalling segmentation, IncenptionV3-based U-Net model outperformed all other algorithms, with an F1-score of 95.66 and 89.43%, respectively. |
|---|---|
| AbstractList | Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity yet increase the accuracy of the damage diagnoses and accessibility in a structural health monitoring (SHM) system. However, despite remarkable advances with DL-based SHM, the most significant challenges to achieving the real-time implication are the limited available defects image databases and the selection of DL networks depth. To address these challenges, this research has created a diverse dataset with concrete crack (4087) and spalling (1100) images and used it for damage condition assessment by applying convolutional neural network (CNN) algorithms. CNN-classifier models are used to identify different types of defects and semantic segmentation for labeling the defect patterns within an image. Three CNN-based models-Visual Geometry Group (VGG)19, ResNet50, and InceptionV3 are incorporated as CNN-classifiers. For semantic segmentation, two encoder-decoder models, U-Net and pyramid scene parsing network architecture are developed based on four backbone models, including VGG19, ResNet50, InceptionV3, and EfficientNetB3. The CNN-classifier models are analyzed on two optimizers-stochastic gradient descent (SGD), root mean square propagation (RMSprop), and learning rates-0.1, 0.001, and 0.0001. However, the CNN-segmentation models are analyzed for SGD and adaptive moment estimation, trained with three different learning rates-0.1, 0.01, and 0.0001, and evaluated based on accuracy, intersection over union, precision, recall, and F1-score. InceptionV3 achieves the best performance for defects classification with an accuracy of 91.98% using the RMSprop optimizer. For crack segmentation, EfficientNetB3-based U-Net, and for spalling segmentation, IncenptionV3-based U-Net model outperformed all other algorithms, with an F1-score of 95.66 and 89.43%, respectively. Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity yet increase the accuracy of the damage diagnoses and accessibility in a structural health monitoring (SHM) system. However, despite remarkable advances with DL-based SHM, the most significant challenges to achieving the real-time implication are the limited available defects image databases and the selection of DL networks depth. To address these challenges, this research has created a diverse dataset with concrete crack (4087) and spalling (1100) images and used it for damage condition assessment by applying convolutional neural network (CNN) algorithms. CNN-classifier models are used to identify different types of defects and semantic segmentation for labeling the defect patterns within an image. Three CNN-based models-Visual Geometry Group (VGG)19, ResNet50, and InceptionV3 are incorporated as CNN-classifiers. For semantic segmentation, two encoder-decoder models, U-Net and pyramid scene parsing network architecture are developed based on four backbone models, including VGG19, ResNet50, InceptionV3, and EfficientNetB3. The CNN-classifier models are analyzed on two optimizers-stochastic gradient descent (SGD), root mean square propagation (RMSprop), and learning rates-0.1, 0.001, and 0.0001. However, the CNN-segmentation models are analyzed for SGD and adaptive moment estimation, trained with three different learning rates-0.1, 0.01, and 0.0001, and evaluated based on accuracy, intersection over union, precision, recall, and F1-score. InceptionV3 achieves the best performance for defects classification with an accuracy of 91.98% using the RMSprop optimizer. For crack segmentation, EfficientNetB3-based U-Net, and for spalling segmentation, IncenptionV3-based U-Net model outperformed all other algorithms, with an F1-score of 95.66 and 89.43%, respectively.Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity yet increase the accuracy of the damage diagnoses and accessibility in a structural health monitoring (SHM) system. However, despite remarkable advances with DL-based SHM, the most significant challenges to achieving the real-time implication are the limited available defects image databases and the selection of DL networks depth. To address these challenges, this research has created a diverse dataset with concrete crack (4087) and spalling (1100) images and used it for damage condition assessment by applying convolutional neural network (CNN) algorithms. CNN-classifier models are used to identify different types of defects and semantic segmentation for labeling the defect patterns within an image. Three CNN-based models-Visual Geometry Group (VGG)19, ResNet50, and InceptionV3 are incorporated as CNN-classifiers. For semantic segmentation, two encoder-decoder models, U-Net and pyramid scene parsing network architecture are developed based on four backbone models, including VGG19, ResNet50, InceptionV3, and EfficientNetB3. The CNN-classifier models are analyzed on two optimizers-stochastic gradient descent (SGD), root mean square propagation (RMSprop), and learning rates-0.1, 0.001, and 0.0001. However, the CNN-segmentation models are analyzed for SGD and adaptive moment estimation, trained with three different learning rates-0.1, 0.01, and 0.0001, and evaluated based on accuracy, intersection over union, precision, recall, and F1-score. InceptionV3 achieves the best performance for defects classification with an accuracy of 91.98% using the RMSprop optimizer. For crack segmentation, EfficientNetB3-based U-Net, and for spalling segmentation, IncenptionV3-based U-Net model outperformed all other algorithms, with an F1-score of 95.66 and 89.43%, respectively. |
| Author | Arafin, Palisa Issa, Anas Billah, Ahm Muntasir |
| Author_xml | – sequence: 1 givenname: Palisa surname: Arafin fullname: Arafin, Palisa organization: Department of Civil Engineering, Lakehead University, Thunder Bay, ON, Canada – sequence: 2 givenname: Ahm Muntasir orcidid: 0000-0001-9840-3438 surname: Billah fullname: Billah, Ahm Muntasir organization: Department of Civil Engineering, University of Calgary, Calgary, AB, Canada – sequence: 3 givenname: Anas surname: Issa fullname: Issa, Anas organization: Department of Civil and Environmental Engineering, United Arab Emirates University, Al Ain, Abu Dhabi, UAE |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/38078049$$D View this record in MEDLINE/PubMed |
| BookMark | eNo1kD9PwzAQxT0U0T_wAVhQRpaA7Th1PKJCAakSC4woOp8vlVHilNgZ-PakUKZ7997v3nBLNgt9IMauBL8VQus7oXRppNCyEGJdSSFnbHH08qM5Z8sYPzmfpF6fs3lRcV1xZRbs44HokLUEQ_Bhn1uI5DLsAw6UKHPUEKaYYQsx-sYjJN-HDIKbojRFx22M02UWqYOQPE5i31FIv-QFO2ugjXR5miv2vn182zznu9enl839LkelTMoRnJGgrHSCNyXHsuDaWmNAWuRacAFkK8mtUqpxoBGdUgJUUwKCtpWTK3bz13sY-q-RYqo7H5HaFgL1Y6yl4dIUleDFhF6f0NF25OrD4DsYvuv_l8gfvFtlWQ |
| CitedBy_id | crossref_primary_10_1299_transjsme_25_00014 crossref_primary_10_3390_jimaging11090288 crossref_primary_10_1007_s11709_024_1048_4 crossref_primary_10_1016_j_eswa_2024_124689 crossref_primary_10_3390_gels10080517 crossref_primary_10_1109_ACCESS_2024_3467989 crossref_primary_10_1016_j_autcon_2025_106230 crossref_primary_10_1016_j_jobe_2025_113589 crossref_primary_10_3390_electronics13050866 crossref_primary_10_1038_s41598_025_87173_7 crossref_primary_10_11648_j_ajce_20251304_12 crossref_primary_10_1007_s11760_025_03913_2 crossref_primary_10_1177_14759217251368998 crossref_primary_10_1016_j_advengsoft_2025_104010 crossref_primary_10_3390_math12193105 crossref_primary_10_1117_1_JEI_33_6_063007 crossref_primary_10_1007_s10845_025_02658_6 crossref_primary_10_1016_j_eswa_2025_128683 crossref_primary_10_1016_j_procs_2024_10_328 crossref_primary_10_1016_j_cscm_2024_e03711 crossref_primary_10_1177_14759217241246953 crossref_primary_10_1016_j_autcon_2025_106349 crossref_primary_10_1007_s10921_024_01103_7 crossref_primary_10_1038_s41598_024_75723_4 crossref_primary_10_1177_09544054241290573 crossref_primary_10_3390_app15189865 crossref_primary_10_1016_j_jobe_2025_114105 crossref_primary_10_1038_s41598_024_82612_3 crossref_primary_10_3390_electronics13163241 crossref_primary_10_1016_j_autcon_2025_106045 crossref_primary_10_3390_buildings14061580 crossref_primary_10_1007_s11227_025_07509_y crossref_primary_10_1007_s00521_024_10962_0 crossref_primary_10_1016_j_dsp_2025_105323 crossref_primary_10_1109_ACCESS_2024_3525183 crossref_primary_10_13168_cs_2024_0025 crossref_primary_10_3390_buildings13112754 crossref_primary_10_1007_s11831_025_10279_8 crossref_primary_10_1177_13694332251381215 crossref_primary_10_3390_drones8070341 crossref_primary_10_7717_peerj_cs_3149 crossref_primary_10_4018_JCIT_349740 |
| ContentType | Journal Article |
| Copyright | The Author(s) 2023. |
| Copyright_xml | – notice: The Author(s) 2023. |
| DBID | NPM 7X8 |
| DOI | 10.1177/14759217231168212 |
| DatabaseName | PubMed MEDLINE - Academic |
| DatabaseTitle | PubMed MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Engineering |
| ExternalDocumentID | 38078049 |
| Genre | Journal Article |
| GroupedDBID | -TM -TN .2N 01A 0R~ 123 1~K 29Q 31X 31Z 4.4 5VS AACTG AADUE AAJPV AAQDB AAQXI AARIX AATAA AATBZ ABAWP ABCCA ABCJG ABEIX ABFWQ ABHQH ABKRH ABLUO ABPNF ABQXT ABRHV ABUJY ACDXX ACGFS ACJER ACLZU ACOFE ACOXC ACROE ACRPL ACSIQ ACTQU ACUAV ACUIR ACXKE ADNMO ADNON ADRRZ ADVBO ADYCS AEDFJ AEPTA AEQLS AESZF AEUHG AEUIJ AEWDL AEWHI AEXNY AFEET AFKRG AFMOU AFQAA AFUIA AFWMB AGKLV AGNHF AGWFA AGWNL AHDMH AHHFK AIOMO AJUZI ALFTD ALMA_UNASSIGNED_HOLDINGS ANDLU ARTOV ASPBG AUTPY AUVAJ AVWKF AYAKG AZFZN B8Z B94 BBRGL BDDNI BPACV CAG CFDXU COF CS3 DH. DO- DOPDO DV7 EBS EJD FEDTE FHBDP GROUPED_SAGE_PREMIER_JOURNAL_COLLECTION H13 HF~ HVGLF HZ~ J8X K.F M4V N9A NPM O9- P.B P2P Q83 RIG ROL S01 SAUOL SCNPE SFC SFK SFS SFT SFX SGV SGZ SPJ SPP SPV STM ZPPRI ZRKOI 7X8 AAPII ABIDT ADDLC ADEBD AJGYC AJHME AJVBE SASJQ |
| ID | FETCH-LOGICAL-c449t-cad92a4b2d10f50c5307bb99a2bc07101aeb820b444fda7ccd441a4f5aca7b8d2 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 51 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000980293500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1475-9217 |
| IngestDate | Sun Nov 09 12:47:59 EST 2025 Wed Feb 19 02:08:07 EST 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 1 |
| Keywords | encoder-decoder model structural health monitoring semantic segmentation convolutional neural network Concrete defects |
| Language | English |
| License | The Author(s) 2023. |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c449t-cad92a4b2d10f50c5307bb99a2bc07101aeb820b444fda7ccd441a4f5aca7b8d2 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0001-9840-3438 |
| OpenAccessLink | https://journals.sagepub.com/doi/full/10.1177/14759217231168212 |
| PMID | 38078049 |
| PQID | 2902938103 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_2902938103 pubmed_primary_38078049 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-01-01 |
| PublicationDateYYYYMMDD | 2024-01-01 |
| PublicationDate_xml | – month: 01 year: 2024 text: 2024-01-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | England |
| PublicationPlace_xml | – name: England |
| PublicationTitle | Structural health monitoring |
| PublicationTitleAlternate | Struct Health Monit |
| PublicationYear | 2024 |
| SSID | ssj0021776 |
| Score | 2.5582592 |
| Snippet | Visual damage detection of infrastructure using deep learning (DL)-based computational approaches can facilitate a potential solution to reduce subjectivity... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| StartPage | 383 |
| Title | Deep learning-based concrete defects classification and detection using semantic segmentation |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/38078049 https://www.proquest.com/docview/2902938103 |
| Volume | 23 |
| WOSCitedRecordID | wos000980293500001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV07T8MwELaAMsDA-1FeMhKrReI6TTwhBFQsVB1A6oKi86vqQFpI4Pdzl6TAgoTEEiVnWYp8Z9_nezJ2ISkLB7IgQNsglIdEgAmRAE8EIC3k6mYT6XCYjcd61BrcyjascnEm1ge1m1mykV9KHaFmyuKodzV_FdQ1iryrbQuNZdbpIZShkK50_OVFQLSdNtlFaSI0frVeTSq4RDQiIbyJ-5mkhpS_Icxa0ww2__uPW2yjxZj8uhGKbbbkix22_qPy4C57vvV-ztuWERNBusxxvBojhqw8d74O8uCWoDXFEtXs41A4HKrq4K2CU8T8hJf-BXkztfgyeWnzmIo99jS4e7y5F22nBWGV0pWw4LQEZaSLo5BENsGdb4zWII0lDBKDNwgVjFIqOEitdYiiQIUELKQmc3KfrRSzwh8yrpyLM69MP2ijwEnIcEovBGVUMDbYLjtfrF2OkkzuCSj87L3Mv1evyw4aBuTzpuRGXpfFx8vM0R9mH7M1icijsZOcsE7AfexP2ar9qKbl21ktIvgcjh4-AXNdybM |
| linkProvider | ProQuest |
| 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=Deep+learning-based+concrete+defects+classification+and+detection+using+semantic+segmentation&rft.jtitle=Structural+health+monitoring&rft.au=Arafin%2C+Palisa&rft.au=Billah%2C+Ahm+Muntasir&rft.au=Issa%2C+Anas&rft.date=2024-01-01&rft.issn=1475-9217&rft.volume=23&rft.issue=1&rft.spage=383&rft_id=info:doi/10.1177%2F14759217231168212&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1475-9217&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1475-9217&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1475-9217&client=summon |