Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
February 5, 2026

Automatic Recognition and Segmentation of Overlapped GPR Target Signatures

View Full Paper
Ask AI
Bookmark
Share

Authors

QRQiuyang RenBeijing Jiaotong UniversityYWYanhui WangQHQuang P. Ha

Discussion

Loading...

Member takes

Implication

Demonstrates improved segmentation of overlapping GPR signatures from bridge scans, indicating enhanced accuracy.

Key Points

  • The aim is to improve the recognition and segmentation of overlapping GPR signatures in civil infrastructure.
  • Proposed a Mask R-CNN based network to capture spatial relationships between GPR signatures.
  • Introduced an improved intersection over union metric accounting for central distance and aspect ratio.
  • Modified Non-Maximum Suppression and enhanced anchor generative mechanism for better detection.
  • The proposed method accurately detects and segments overlapping GPR signatures.
  • Achieved an average accuracy of 46.8% in the segmentation task.
  • Significantly outperformed existing Mask R-CNN models in terms of segmentation effectiveness.

Cite This Study

Ren et al. (2025) studied this question.

synapsesocial.com/papers/698433c8f1d9ada3c1fb1287https://doi.org/10.1051/e3sconf/202562603003/pdf
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1ROI-Binarized Hyperbolic Region Segmentation and Characterization by Using Deep Residual Convolutional Neural Network with Skip Connection for GPR Imaging2024 · 2 citations
  2. 2Underground target detection method based on GPR physical mechanism constraints2026
  3. 3A Novel Automatic Detection and Positioning Strategy for Buried Cylindrical Objects Based on B-Scan GPR Images2025
  4. 4GPRNet: A Geometric Prior-Refined Semantic Segmentation Network for Land Use and Land Cover Mapping2025 · 4 citations
  5. 5A Generalizable Deep Learning Approach to Classifying Ground Radar Images2026