EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep Learning and Geometric Modeling for 3D Reconstruction of Subsurface Utilities from GPR Data

Peyman Jafary iD, Davood Shojaei iD, Krista A Ehinger iD

Sensors · 17.10.2025 · 25 (20) · 6414

DOI: 10.3390/s25206414 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 17.10.2025. Verfasst von Peyman Jafary und 2 weiteren Autor:innen. Beteiligte Einrichtung: Centre for Spatial Data Infrastructures and Land Administration, Department of Infrastructure Engineering, The University of Melbourne, Melbourne, VI…. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The University of Melbourne, University of Melbourne.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate underground utility mapping remains a critical yet complex task in Ground Penetrating Radar (GPR) interpretation, essential to avoiding costly and dangerous excavation errors. This study presents a novel deep learning-based pipeline for 3D reconstruction of buried linear utilities from high-resolution GPR B-scan data. Three state-of-the-art models—YOLOv8, YOLOv11, and Mask R-CNN—were employed for both bounding box and keypoint detection of hyperbolic reflections, using a real-world GPR dataset. On the test set, Mask R-CNN achieved the highest keypoint F1-score (0.822) and bounding box F1-score (0.867), outperforming the YOLO models. Detected summit points were clustered using a 3D DBSCAN algorithm to approximate the spatial trajectories of buried utilities. RANSAC-based line fitting was then applied to each cluster, yielding an average RMSE of 0.06 across all fitted 3D paths. The key innovation of this hybrid model lies in its use of real-world data (avoiding synthetic augmentation), direct summit point detection (beyond bounding box analysis), and a geometric 3D reconstruction pipeline. This approach addresses key limitations in prior studies, including poor generalizability to complex real-world scenarios and the reliance on full 3D data volumes. Our method offers a more practical and scalable solution for subsurface utility mapping in real-world settings.

Schlagwörter

Stichwörter der Autor:innen

3D utility mappingDBSCANGround Penetrating Radar (GPR)Mask R-CNNRANSACYOLOv11YOLOv8keypoint detection

Autor:innen und Institutionen

  • Centre for Spatial Data Infrastructures and Land Administration, Department of Infrastructure Engineering, The University of Melbourne, Melbourne, VIC 3010, Australia.
  • School of Computing and Information Systems, The University of Melbourne, Melbourne, VIC 3010, Australia.

ORCID-Profile: Peyman Jafary, Davood Shojaei, Krista A Ehinger

Förderung

  • The University of Melbourne
  • University of Melbourne

Zitieren

APA

Jafary, P., Shojaei, D., & Ehinger, K. A. (2025). Deep Learning and Geometric Modeling for 3D Reconstruction of Subsurface Utilities from GPR Data. Sensors, 25(20), 6414. https://doi.org/10.3390/s25206414

Vancouver

Jafary P, Shojaei D, Ehinger KA. Deep Learning and Geometric Modeling for 3D Reconstruction of Subsurface Utilities from GPR Data. Sensors. 2025;25(20):6414. doi:10.3390/s25206414

BibTeX
@article{Jafary2025,
  author = {Jafary, Peyman and Shojaei, Davood and Ehinger, Krista A},
  title = {Deep Learning and Geometric Modeling for 3D Reconstruction of Subsurface Utilities from GPR Data},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {20},
  pages = {6414},
  doi = {10.3390/s25206414}
}
Kostenlos starten →