IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems

Deep learning has recently been extensively used for crack detection in structural health monitoring settings. However, detecting cracks in levee systems have yet to receive considerable critical attention. Thus, this study presents a novel encoder-decoder-based fully convolutional neural network to...

Full description

Saved in:
Bibliographic Details
Main Authors: Manisha Panta, Md Tamjidul Hoque, Mahdi Abdelguerfi, Maik C. Flanagin
Format: Article
Language:English
Published: IEEE 2023-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10035954/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846128528001597440
author Manisha Panta
Md Tamjidul Hoque
Mahdi Abdelguerfi
Maik C. Flanagin
author_facet Manisha Panta
Md Tamjidul Hoque
Mahdi Abdelguerfi
Maik C. Flanagin
author_sort Manisha Panta
collection DOAJ
description Deep learning has recently been extensively used for crack detection in structural health monitoring settings. However, detecting cracks in levee systems have yet to receive considerable critical attention. Thus, this study presents a novel encoder-decoder-based fully convolutional neural network to detect cracks from levee images at a pixel level automatically. We propose that the feature learning be strengthened using the decoder and bottleneck feature maps by concatenating them back to the encoder blocks. The addition reinforcement in the U-Net-like architecture results in a loop-like structure to exploit all the feature maps from encoders, bottlenecks, and decoders. The proposed architecture, Iterative Loop U-Net (IterLUNet), outperforms the state-of-the-art architectures on the image dataset of the levee system, achieving an increment of Intersection over Union (IoU) by 10.32% on average for a 10-Fold Cross-Validation (FCV) compared to the baseline U-Net model and 11.00%, 7.65%, and 7.43% with a range of latest models MultiResUnet, Attention U-Net, and Unet++ respectively. In addition, IterLUNet has at least 63% fewer parameters to be trained than the baseline model, thus, allowing less space consumption for pixel-wise crack detection in AI-based inspection of levee systems.
format Article
id doaj-art-3d9d0e1f4d0e4dbcbb395014907846b8
institution Kabale University
issn 2169-3536
language English
publishDate 2023-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-3d9d0e1f4d0e4dbcbb395014907846b82024-12-11T00:04:04ZengIEEEIEEE Access2169-35362023-01-0111122491226210.1109/ACCESS.2023.324187710035954IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee SystemsManisha Panta0Md Tamjidul Hoque1https://orcid.org/0000-0002-0110-2194Mahdi Abdelguerfi2Maik C. Flanagin3Canizaro Livingston Gulf States Center for Environmental Informatics, The University of New Orleans, New Orleans, LA, USACanizaro Livingston Gulf States Center for Environmental Informatics, The University of New Orleans, New Orleans, LA, USACanizaro Livingston Gulf States Center for Environmental Informatics, The University of New Orleans, New Orleans, LA, USAU.S. Army Corps of Engineers, New Orleans, LA, USADeep learning has recently been extensively used for crack detection in structural health monitoring settings. However, detecting cracks in levee systems have yet to receive considerable critical attention. Thus, this study presents a novel encoder-decoder-based fully convolutional neural network to detect cracks from levee images at a pixel level automatically. We propose that the feature learning be strengthened using the decoder and bottleneck feature maps by concatenating them back to the encoder blocks. The addition reinforcement in the U-Net-like architecture results in a loop-like structure to exploit all the feature maps from encoders, bottlenecks, and decoders. The proposed architecture, Iterative Loop U-Net (IterLUNet), outperforms the state-of-the-art architectures on the image dataset of the levee system, achieving an increment of Intersection over Union (IoU) by 10.32% on average for a 10-Fold Cross-Validation (FCV) compared to the baseline U-Net model and 11.00%, 7.65%, and 7.43% with a range of latest models MultiResUnet, Attention U-Net, and Unet++ respectively. In addition, IterLUNet has at least 63% fewer parameters to be trained than the baseline model, thus, allowing less space consumption for pixel-wise crack detection in AI-based inspection of levee systems.https://ieeexplore.ieee.org/document/10035954/Crack detectiondeep learningfloodwallsimage segmentationlevees
spellingShingle Manisha Panta
Md Tamjidul Hoque
Mahdi Abdelguerfi
Maik C. Flanagin
IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
IEEE Access
Crack detection
deep learning
floodwalls
image segmentation
levees
title IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
title_full IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
title_fullStr IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
title_full_unstemmed IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
title_short IterLUNet: Deep Learning Architecture for Pixel-Wise Crack Detection in Levee Systems
title_sort iterlunet deep learning architecture for pixel wise crack detection in levee systems
topic Crack detection
deep learning
floodwalls
image segmentation
levees
url https://ieeexplore.ieee.org/document/10035954/
work_keys_str_mv AT manishapanta iterlunetdeeplearningarchitectureforpixelwisecrackdetectioninleveesystems
AT mdtamjidulhoque iterlunetdeeplearningarchitectureforpixelwisecrackdetectioninleveesystems
AT mahdiabdelguerfi iterlunetdeeplearningarchitectureforpixelwisecrackdetectioninleveesystems
AT maikcflanagin iterlunetdeeplearningarchitectureforpixelwisecrackdetectioninleveesystems