A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level
Cracks are the most prevalent form of damage on pavement surfaces. Accurately recognizing pavement cracks is often difficult due to background interference and other challenges. Moreover, accurate and fast automatic detection of road cracks plays a crucial role in assessing pavement conditions. Ther...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10716393/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841546154670555136 |
|---|---|
| author | Zhong Luo Xinle Li Yanfeng Zheng |
| author_facet | Zhong Luo Xinle Li Yanfeng Zheng |
| author_sort | Zhong Luo |
| collection | DOAJ |
| description | Cracks are the most prevalent form of damage on pavement surfaces. Accurately recognizing pavement cracks is often difficult due to background interference and other challenges. Moreover, accurate and fast automatic detection of road cracks plays a crucial role in assessing pavement conditions. Therefore, a highly efficient lightweight network with only 0.78M parameters is proposed for the pixel-level pavement crack detection task. In this paper, adaptive enhancement module (AEM) is designed and added to the encoder network in order to avoid the problem of insufficient model learning capability due to the use of depth-wise separable convolutions. Meanwhile, a coordinate-aware fusion module (CFM) is proposed, which fully fuses skip connection features and decoder features to enhance cross-channel interaction information. Comprehensive experimental results demonstrate that the proposed network outperforms existing methods across four public datasets: CamCrack789, CFD, DeepCrack237, and Crack500, achieving F1 scores of 94.6%, 92.8%, 91.0%, and 79.8%, respectively. Furthermore, ablation study confirmed the efficacy of both the AEM and the CFM. |
| format | Article |
| id | doaj-art-b5d3048e1c514b408b1cfbc0ba45fb98 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-b5d3048e1c514b408b1cfbc0ba45fb982025-01-11T00:00:41ZengIEEEIEEE Access2169-35362024-01-011215338515339410.1109/ACCESS.2024.347924510716393A Novel Lightweight U-Shaped Network for Crack Detection at Pixel LevelZhong Luo0https://orcid.org/0009-0003-2581-9266Xinle Li1https://orcid.org/0000-0003-0490-2020Yanfeng Zheng2https://orcid.org/0009-0009-2829-5641College of Civil Engineering, Dalian Minzu University, Dalian, Liaoning, ChinaCollege of Civil Engineering, Dalian Minzu University, Dalian, Liaoning, ChinaDepartment of Computer Science, Dalian Minzu University, Dalian, Liaoning, ChinaCracks are the most prevalent form of damage on pavement surfaces. Accurately recognizing pavement cracks is often difficult due to background interference and other challenges. Moreover, accurate and fast automatic detection of road cracks plays a crucial role in assessing pavement conditions. Therefore, a highly efficient lightweight network with only 0.78M parameters is proposed for the pixel-level pavement crack detection task. In this paper, adaptive enhancement module (AEM) is designed and added to the encoder network in order to avoid the problem of insufficient model learning capability due to the use of depth-wise separable convolutions. Meanwhile, a coordinate-aware fusion module (CFM) is proposed, which fully fuses skip connection features and decoder features to enhance cross-channel interaction information. Comprehensive experimental results demonstrate that the proposed network outperforms existing methods across four public datasets: CamCrack789, CFD, DeepCrack237, and Crack500, achieving F1 scores of 94.6%, 92.8%, 91.0%, and 79.8%, respectively. Furthermore, ablation study confirmed the efficacy of both the AEM and the CFM.https://ieeexplore.ieee.org/document/10716393/Crack detectionpixel-leveldeep learningconvolution neural networks |
| spellingShingle | Zhong Luo Xinle Li Yanfeng Zheng A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level IEEE Access Crack detection pixel-level deep learning convolution neural networks |
| title | A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level |
| title_full | A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level |
| title_fullStr | A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level |
| title_full_unstemmed | A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level |
| title_short | A Novel Lightweight U-Shaped Network for Crack Detection at Pixel Level |
| title_sort | novel lightweight u shaped network for crack detection at pixel level |
| topic | Crack detection pixel-level deep learning convolution neural networks |
| url | https://ieeexplore.ieee.org/document/10716393/ |
| work_keys_str_mv | AT zhongluo anovellightweightushapednetworkforcrackdetectionatpixellevel AT xinleli anovellightweightushapednetworkforcrackdetectionatpixellevel AT yanfengzheng anovellightweightushapednetworkforcrackdetectionatpixellevel AT zhongluo novellightweightushapednetworkforcrackdetectionatpixellevel AT xinleli novellightweightushapednetworkforcrackdetectionatpixellevel AT yanfengzheng novellightweightushapednetworkforcrackdetectionatpixellevel |