ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation

Accurate segmentation of pavement cracks from high-resolution remote sensing imagery plays a crucial role in automated road condition assessment and infrastructure maintenance. However, crack structures often exhibit asymmetry, irregular morphology, and multi-scale variations, posing significant cha...

Full description

Saved in:
Bibliographic Details
Main Authors: Chao Tan, Jiaqi Liu, Zhedong Zhao, Rufei Liu, Peng Tan, Aishu Yao, Shoudao Pan, Jingyi Dong
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/11/6183
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850129027261530112
author Chao Tan
Jiaqi Liu
Zhedong Zhao
Rufei Liu
Peng Tan
Aishu Yao
Shoudao Pan
Jingyi Dong
author_facet Chao Tan
Jiaqi Liu
Zhedong Zhao
Rufei Liu
Peng Tan
Aishu Yao
Shoudao Pan
Jingyi Dong
author_sort Chao Tan
collection DOAJ
description Accurate segmentation of pavement cracks from high-resolution remote sensing imagery plays a crucial role in automated road condition assessment and infrastructure maintenance. However, crack structures often exhibit asymmetry, irregular morphology, and multi-scale variations, posing significant challenges to conventional CNN-based methods in real-world environments. Specifically, the proposed ETAFHrNet focuses on two predominant pavement-distress morphologies—linear cracks (transverse and longitudinal) and alligator cracks—and has been empirically validated on their intersections and branching patterns over both asphalt and concrete road surfaces. In this work, we present ETAFHrNet, a novel attention-guided segmentation network designed to address the limitations of traditional architectures in detecting fine-grained and asymmetric patterns. ETAFHrNet integrates Transformer-based global attention and multi-scale hybrid feature fusion, enhancing both contextual perception and detail sensitivity. The network introduces two key modules: the Efficient Hybrid Attention Transformer (EHAT), which captures long-range dependencies, and the Cross-Scale Hybrid Attention Module (CSHAM), which adaptively fuses features across spatial resolutions. To support model training and benchmarking, we also propose QD-Crack, a high-resolution, pixel-level annotated dataset collected from real-world road inspection scenarios. Experimental results show that ETAFHrNet significantly outperforms existing methods—including U-Net, DeepLabv3+, and HRNet—in both segmentation accuracy and generalization ability. These findings demonstrate the effectiveness of interpretable, multi-scale attention architectures in complex object detection and image classification tasks, making our approach relevant for broader applications, such as autonomous driving, remote sensing, and smart infrastructure systems.
format Article
id doaj-art-a6b204ceeb704995920bca26c854f070
institution OA Journals
issn 2076-3417
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-a6b204ceeb704995920bca26c854f0702025-08-20T02:33:07ZengMDPI AGApplied Sciences2076-34172025-05-011511618310.3390/app15116183ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack SegmentationChao Tan0Jiaqi Liu1Zhedong Zhao2Rufei Liu3Peng Tan4Aishu Yao5Shoudao Pan6Jingyi Dong7College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaCollege of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, ChinaAccurate segmentation of pavement cracks from high-resolution remote sensing imagery plays a crucial role in automated road condition assessment and infrastructure maintenance. However, crack structures often exhibit asymmetry, irregular morphology, and multi-scale variations, posing significant challenges to conventional CNN-based methods in real-world environments. Specifically, the proposed ETAFHrNet focuses on two predominant pavement-distress morphologies—linear cracks (transverse and longitudinal) and alligator cracks—and has been empirically validated on their intersections and branching patterns over both asphalt and concrete road surfaces. In this work, we present ETAFHrNet, a novel attention-guided segmentation network designed to address the limitations of traditional architectures in detecting fine-grained and asymmetric patterns. ETAFHrNet integrates Transformer-based global attention and multi-scale hybrid feature fusion, enhancing both contextual perception and detail sensitivity. The network introduces two key modules: the Efficient Hybrid Attention Transformer (EHAT), which captures long-range dependencies, and the Cross-Scale Hybrid Attention Module (CSHAM), which adaptively fuses features across spatial resolutions. To support model training and benchmarking, we also propose QD-Crack, a high-resolution, pixel-level annotated dataset collected from real-world road inspection scenarios. Experimental results show that ETAFHrNet significantly outperforms existing methods—including U-Net, DeepLabv3+, and HRNet—in both segmentation accuracy and generalization ability. These findings demonstrate the effectiveness of interpretable, multi-scale attention architectures in complex object detection and image classification tasks, making our approach relevant for broader applications, such as autonomous driving, remote sensing, and smart infrastructure systems.https://www.mdpi.com/2076-3417/15/11/6183pavement crack segmentationtransformer neural networksmulti-scale feature fusionglobal attention mechanismhigh-resolution remote sensingdeep learning
spellingShingle Chao Tan
Jiaqi Liu
Zhedong Zhao
Rufei Liu
Peng Tan
Aishu Yao
Shoudao Pan
Jingyi Dong
ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
Applied Sciences
pavement crack segmentation
transformer neural networks
multi-scale feature fusion
global attention mechanism
high-resolution remote sensing
deep learning
title ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
title_full ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
title_fullStr ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
title_full_unstemmed ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
title_short ETAFHrNet: A Transformer-Based Multi-Scale Network for Asymmetric Pavement Crack Segmentation
title_sort etafhrnet a transformer based multi scale network for asymmetric pavement crack segmentation
topic pavement crack segmentation
transformer neural networks
multi-scale feature fusion
global attention mechanism
high-resolution remote sensing
deep learning
url https://www.mdpi.com/2076-3417/15/11/6183
work_keys_str_mv AT chaotan etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT jiaqiliu etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT zhedongzhao etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT rufeiliu etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT pengtan etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT aishuyao etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT shoudaopan etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation
AT jingyidong etafhrnetatransformerbasedmultiscalenetworkforasymmetricpavementcracksegmentation