EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Transformer–CNN Hybrid Framework for Pavement Pothole Segmentation

Tianjie Zhang, Zhen Liu iD, Bingyan Cui iD, Xingyu Gu iD, Yang Lu iD

Sensors · 04.11.2025 · 25 (21) · 6756

DOI: 10.3390/s25216756 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 04.11.2025. Verfasst von Tianjie Zhang und 4 weiteren Autor:innen. Beteiligte Einrichtung: Center for Advanced Infrastructure and Transportation, Rutgers University, Piscataway, NJ 08854, USA. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Pavement surface defects such as potholes pose significant safety risks and accelerate infrastructure deterioration. Accurate and automated detection of such defects requires both advanced sensing technologies and robust deep learning models. In this study, we propose PoFormer, a Transformer–CNN hybrid framework designed for precise segmentation of pavement potholes from heterogeneous image datasets. The architecture leverages the global feature extraction ability of Transformers and the fine-grained localization capability of CNNs, achieving superior segmentation accuracy compared to state-of-the-art models. To construct a representative dataset, we combined open source images with high-resolution field data acquired using a multi-sensor pavement inspection vehicle equipped with a line-scan camera and infrared/laser-assisted lighting. This sensing system provides millimeter-level resolution and continuous 3D surface imaging under diverse environmental conditions, ensuring robust training inputs for deep learning. Experimental results demonstrate that PoFormer achieves a mean IoU of 77.23% and a mean pixel accuracy of 84.48%, outperforming existing CNN-based models. By integrating multi-sensor data acquisition with advanced hybrid neural networks, this work highlights the potential of 3D imaging and sensing technologies for intelligent pavement condition monitoring and automated infrastructure maintenance.

Schlagwörter

Stichwörter der Autor:innen

CNNdeep learningimage segmentationpotholetransformer

Autor:innen und Institutionen

  • Center for Advanced Infrastructure and Transportation, Rutgers University, Piscataway, NJ 08854, USA.
  • Department of Engineering, Boise State University, Boise, ID 83752, USA.
  • Institute of Space and Earth Information Science, Fok Ying Tung Remote Sensing Science Building, The Chinese University of Hong Kong, Hong Kong SAR, China.
  • School of Transportation, Southeast University, Nanjing 211189, China.

ORCID-Profile: Zhen Liu, Bingyan Cui, Xingyu Gu, Yang Lu

Zitieren

APA

Zhang, T., Liu, Z., Cui, B., Gu, X., & Lu, Y. (2025). Transformer–CNN Hybrid Framework for Pavement Pothole Segmentation. Sensors, 25(21), 6756. https://doi.org/10.3390/s25216756

Vancouver

Zhang T, Liu Z, Cui B, Gu X, Lu Y. Transformer–CNN Hybrid Framework for Pavement Pothole Segmentation. Sensors. 2025;25(21):6756. doi:10.3390/s25216756

BibTeX
@article{Zhang2025,
  author = {Zhang, Tianjie and Liu, Zhen and Cui, Bingyan and Gu, Xingyu and Lu, Yang},
  title = {Transformer–CNN Hybrid Framework for Pavement Pothole Segmentation},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {21},
  pages = {6756},
  doi = {10.3390/s25216756}
}
Kostenlos starten →