Segmentation Tool for Images of Cracks

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

2 Citaten (Scopus)
4 Downloads (Pure)

Samenvatting

Safety-critical infrastructures, such as bridges, are periodically inspected to check for existing damage, such as fatigue cracks and corrosion, and to guarantee the safe use of the infrastructure. Visual inspection is the most frequent type of general inspection, despite the fact that its detection capability is rather limited, especially for fatigue cracks. Machine learning algorithms can be used for augmenting the capability of classical visual inspection of bridge structures, however, the implementation of such an algorithm requires a massive annotated training dataset, which is time-consuming to produce. This paper proposes a semi-automatic crack segmentation tool that eases the manual segmentation of cracks on images needed to create a training dataset for machine learning algorithm. Also it can be used to measure the geometry of the crack. This tool makes use of an image processing algorithm, which was initially developed for the analysis of vascular systems on retinal images. The algorithm relies on a multi-orientation wavelet transform, which is applied to the image to construct the so-called ‘orientation scores’, i.e. a modified version of the image. Afterwards, the filtered orientation scores are used to formulate an optimal path problem that identifies the crack. The globally optimal path between manually selected crack endpoints is computed, using a state-of-the-art geometric tracking method. The pixel-wise segmentation is done afterwards using the obtained crack path. The proposed method outperforms fully automatic methods and shows potential to be an adequate alternative to the manual data annotation.
Originele taal-2Engels
TitelAdvances in Information Technology in Civil and Building Engineering
SubtitelProceedings of ICCCBE 2022 - Volume 1
RedacteurenSebastian Skatulla, Hans Beushausen
Plaats van productieCham
UitgeverijSpringer
Pagina's93-110
Aantal pagina's18
ISBN van elektronische versie978-3-031-35399-4
ISBN van geprinte versie978-3-031-35401-4, 978-3-031-35398-7
DOI's
StatusGepubliceerd - 30 sep. 2023
Evenement19th International Conference on Computing in Civil and Building Engineering, ICCCBE 2022 - Cape Town, Zuid-Afrika
Duur: 26 okt. 202228 okt. 2022

Publicatie series

NaamLecture Notes in Civil Engineering (LNCE)
Volume357
ISSN van geprinte versie2366-2557
ISSN van elektronische versie2366-2565

Congres

Congres19th International Conference on Computing in Civil and Building Engineering, ICCCBE 2022
Land/RegioZuid-Afrika
StadCape Town
Periode26/10/2228/10/22

Financiering

Authors would like to thank the Dutch bridge infrastructure owners “ProRail” and “Rijkswaterstaat” for their support. The research is primarily funded by the Eindhoven Artificial Intelligence Systems Institute, and partly by the Dutch Foundation of Science NWO (Geometric learning for Image Analysis, VI.C 202-031).

FinanciersFinanciernummer
Nederlandse Organisatie voor Wetenschappelijk OnderzoekVI.C 202-031
Eindhoven University of Technology

    Vingerafdruk

    Duik in de onderzoeksthema's van 'Segmentation Tool for Images of Cracks'. Samen vormen ze een unieke vingerafdruk.

    Citeer dit