Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms

Publikation: KonferenzbeitragPaperForschung

Autoren

Organisationseinheiten

Externe Organisationen

  • Hochschule Konstanz Technik, Wirtschaft und Gestaltung (HTWG)
  • Dr. Hesse und Partner Ingenieure
  • WKC Hamburg GmbH
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
PublikationsstatusElektronisch veröffentlicht (E-Pub) - 2020
VeranstaltungFIG Working Week 2020: Smart surveyors for land and water management - Amsterdam, Niederlande
Dauer: 10 Mai 202014 Mai 2020

Konferenz

KonferenzFIG Working Week 2020
Land/GebietNiederlande
OrtAmsterdam
Zeitraum10 Mai 202014 Mai 2020

Abstract

The ageing infrastructure in ports requires regular inspection. This inspection is currently carried out manually by divers who sense by hand the entire underwater infrastructure. This process is cost-intensive as it involves a lot of time and human resources. To overcome these difficulties, we propose to scan the above and underwater port structure with a Multi-Sensor-System, and -by a fully automated process-to classify the obtained point cloud into damaged and undamaged zones. We make use of simulated training data to test our approach since not enough training data with corresponding class labels are available yet. To that aim, we build a rasterised heightfield of a point cloud of a sheet pile wall by cutting it into verticall slices. The distance from each slice to the corresponding line generates the heightfield. This latter is propagated through a convolutional neural network which detects anomalies. We use the VGG19 Deep Neural Network model pretrained on natural images. This neural network has 19 layers and it is often used for image recognition tasks. We showed that our approach can achieve a fully automated, reproducible, quality-controlled damage detection which is able to analyse the whole structure instead of the sample wise manual method with divers. The mean true positive rate is 0.98 which means that we detected 98 % of the damages in the simulated environment.

Zitieren

Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms. / Hake, Frederic; Hermann, Matthias; Alkhatib, Hamza et al.
2020. Beitrag in FIG Working Week 2020, Amsterdam, Niederlande.

Publikation: KonferenzbeitragPaperForschung

Hake F, Hermann M, Alkhatib H, Hesse C, Holste K, Umlauf G et al.. Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms. 2020. Beitrag in FIG Working Week 2020, Amsterdam, Niederlande. Epub 2020.
Hake, Frederic ; Hermann, Matthias ; Alkhatib, Hamza et al. / Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms. Beitrag in FIG Working Week 2020, Amsterdam, Niederlande.
Download
@conference{3c1481c1a4c54a35b099bcd44ed45c7d,
title = "Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms",
abstract = "The ageing infrastructure in ports requires regular inspection. This inspection is currently carried out manually by divers who sense by hand the entire underwater infrastructure. This process is cost-intensive as it involves a lot of time and human resources. To overcome these difficulties, we propose to scan the above and underwater port structure with a Multi-Sensor-System, and -by a fully automated process-to classify the obtained point cloud into damaged and undamaged zones. We make use of simulated training data to test our approach since not enough training data with corresponding class labels are available yet. To that aim, we build a rasterised heightfield of a point cloud of a sheet pile wall by cutting it into verticall slices. The distance from each slice to the corresponding line generates the heightfield. This latter is propagated through a convolutional neural network which detects anomalies. We use the VGG19 Deep Neural Network model pretrained on natural images. This neural network has 19 layers and it is often used for image recognition tasks. We showed that our approach can achieve a fully automated, reproducible, quality-controlled damage detection which is able to analyse the whole structure instead of the sample wise manual method with divers. The mean true positive rate is 0.98 which means that we detected 98 % of the damages in the simulated environment.",
author = "Frederic Hake and Matthias Hermann and Hamza Alkhatib and Christian Hesse and Karsten Holste and Georg Umlauf and Gael Kermarrec and Ingo Neumann",
year = "2020",
language = "English",
note = "FIG Working Week 2020 ; Conference date: 10-05-2020 Through 14-05-2020",

}

Download

TY - CONF

T1 - Damage Detection for Port Infrastructure by Means of Machine-Learning-Algorithms

AU - Hake, Frederic

AU - Hermann, Matthias

AU - Alkhatib, Hamza

AU - Hesse, Christian

AU - Holste, Karsten

AU - Umlauf, Georg

AU - Kermarrec, Gael

AU - Neumann, Ingo

PY - 2020

Y1 - 2020

N2 - The ageing infrastructure in ports requires regular inspection. This inspection is currently carried out manually by divers who sense by hand the entire underwater infrastructure. This process is cost-intensive as it involves a lot of time and human resources. To overcome these difficulties, we propose to scan the above and underwater port structure with a Multi-Sensor-System, and -by a fully automated process-to classify the obtained point cloud into damaged and undamaged zones. We make use of simulated training data to test our approach since not enough training data with corresponding class labels are available yet. To that aim, we build a rasterised heightfield of a point cloud of a sheet pile wall by cutting it into verticall slices. The distance from each slice to the corresponding line generates the heightfield. This latter is propagated through a convolutional neural network which detects anomalies. We use the VGG19 Deep Neural Network model pretrained on natural images. This neural network has 19 layers and it is often used for image recognition tasks. We showed that our approach can achieve a fully automated, reproducible, quality-controlled damage detection which is able to analyse the whole structure instead of the sample wise manual method with divers. The mean true positive rate is 0.98 which means that we detected 98 % of the damages in the simulated environment.

AB - The ageing infrastructure in ports requires regular inspection. This inspection is currently carried out manually by divers who sense by hand the entire underwater infrastructure. This process is cost-intensive as it involves a lot of time and human resources. To overcome these difficulties, we propose to scan the above and underwater port structure with a Multi-Sensor-System, and -by a fully automated process-to classify the obtained point cloud into damaged and undamaged zones. We make use of simulated training data to test our approach since not enough training data with corresponding class labels are available yet. To that aim, we build a rasterised heightfield of a point cloud of a sheet pile wall by cutting it into verticall slices. The distance from each slice to the corresponding line generates the heightfield. This latter is propagated through a convolutional neural network which detects anomalies. We use the VGG19 Deep Neural Network model pretrained on natural images. This neural network has 19 layers and it is often used for image recognition tasks. We showed that our approach can achieve a fully automated, reproducible, quality-controlled damage detection which is able to analyse the whole structure instead of the sample wise manual method with divers. The mean true positive rate is 0.98 which means that we detected 98 % of the damages in the simulated environment.

M3 - Paper

T2 - FIG Working Week 2020

Y2 - 10 May 2020 through 14 May 2020

ER -

Von denselben Autoren