Skip to content
DraconiticResearch
Search Sources API
203,948 studies31 sourcesUpdated 44 min ago
Peer-reviewedCC BY 4.0PLOS One

Yolov8-CA-EYHL: A bridge crack detection algorithm based on improved YOLOv8

  • Xianwei Zhu1
  • He Chao2
  • Yahui Zhang3
  1. 1School of Mechanical and Electrical Engineering, Zhengzhou University of Industrial Technology, Zhengzhou, China
  2. 2School of Intelligent Manufacturing and Electrical Engineering, Nanyang Normal University, Henan, China
  3. 3School of Intelligent Manufacturing, Nanyang Institute of Technology, Henan, China
Open at sourcePDFDOICite

Abstract

Automated bridge crack detection is challenging because cracks often exhibit weak contrast, irregular morphology, slender structures, and strong interference from complex surface textures. To address these issues, this study proposes a YOLOv8-CA-EYHL framework that combines filtering-equalization preprocessing with coordinate attention (CA) and enhanced high- and low-frequency feature learning modules. The filtering-equalization strategy improves crack-background contrast while suppressing background interference. CA enhances spatially directional feature representation, while the YHL module strengthens local and high-frequency crack features. The EYHL module further establishes multi-scale feature interaction and spatial recalibration to improve the representation of irregular and slender cracks. Ablation experiments demonstrate that the performance improvement results from the complementary contributions of preprocessing and the proposed architectural components. The proposed model achieves an mAP50 of 93.1% and an mAP50-95 of 63.5%, compared with 64.4% and 42.8%, respectively, for YOLOv8s. It requires 32.6 B FLOPs and 13.7 M parameters while maintaining an inference speed of approximately 76 FPS. Repeated experiments with different random seeds and evaluation on an independent external dataset provide further evidence of the stability and generalisation potential of the proposed framework.

Related

  1. CPD-YOLOv8: A Causality-Driven Physical Modality Fusion Framework for Early Micro-Fire Detection to Protect Complex Wildland Vegetation
    Juan Liu, Ruitong Yao, Qingyuan Yang +7
    Plants (Basel, Switzerland)Peer-reviewedCC BYFull textPlant biologyEurope PMC
  2. A method for recognizing and detecting the physiological state of silkworms based on an improved YOLOv8 approach
    Pu Liu, Xijie Liao, Chunhua Liu +3
    Scientific ReportsPeer-reviewedCC BYFull text1 citedPlant biologyDOAJ
  3. Multi-Task Bacterial Colony Detection and Classification Using YOLOv8 with Edge Optimization for Resource-Constrained Deployment
    Belaguppa Manjunath Ashwin Desai, Rohan Rajesh, Shreyas Murthy +2
    arXivPreprintCC BY-NC-SA 4.0Full textBioinformatics & computational biologyarXiv
  4. Integrating YOLOv8-based disaster detection with deep reinforcement learning for closed-loop UAV emergency response scheduling: A simulation study
    Chunsheng Zhang, Ruichun He
    PLOS OnePeer-reviewedCC BYFull textNeuroscienceStatisticsPLOSEurope PMCPMC
  5. An offline deep learning model for bovine muzzle identification using the variance of the laplacian, YOLOv8, EfficientNetV2-S, and ArcFace.
    Ahlam G Khalifa, Ahmed S Abdultawab
    BMC veterinary researchPeer-reviewedCC BYFull textPlant biologyEurope PMC

Cite

BibTeXRIS
BibTeX
@article{zhu2026yolov8caeyhl,
  title = {{Yolov8-CA-EYHL: A bridge crack detection algorithm based on improved YOLOv8}},
  author = {Xianwei Zhu and He Chao and Yahui Zhang},
  journal = {PLOS One},
  year = {2026},
  month = sep,
  doi = {10.1371/journal.pone.0358872},
  url = {https://doi.org/10.1371/journal.pone.0358872},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Yolov8-CA-EYHL: A bridge crack detection algorithm based on improved YOLOv8
AU  - Xianwei Zhu
AU  - He Chao
AU  - Yahui Zhang
PY  - 2026
DA  - 2026/09/22
JO  - PLOS One
DO  - 10.1371/journal.pone.0358872
UR  - https://doi.org/10.1371/journal.pone.0358872
AB  - Automated bridge crack detection is challenging because cracks often exhibit weak contrast, irregular morphology, slender structures, and strong interference from complex surface textures. To address these issues, this study proposes a YOLOv8-CA-EYHL framework that combines filtering-equalization preprocessing with coordinate attention (CA) and enhanced high- and low-frequency feature learning modules. The filtering-equalization strategy improves crack-background contrast while suppressing background interference. CA enhances spatially directional feature representation, while the YHL module strengthens local and high-frequency crack features. The EYHL module further establishes multi-scale feature interaction and spatial recalibration to improve the representation of irregular and slender cracks. Ablation experiments demonstrate that the performance improvement results from the complementary contributions of preprocessing and the proposed architectural components. The proposed model achieves an mAP50 of 93.1% and an mAP50-95 of 63.5%, compared with 64.4% and 42.8%, respectively, for YOLOv8s. It requires 32.6 B FLOPs and 13.7 M parameters while maintaining an inference speed of approximately 76 FPS. Repeated experiments with different random seeds and evaluation on an independent external dataset provide further evidence of the stability and generalisation potential of the proposed framework.
SN  - 1932-6203
ER  -