Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
February 5, 2026Applied SciencesOpen Access

GA-HRNet: High-Precision Building Extraction for Individualization of Oblique Photogrammetry 3D Models

View Full Paper
Ask AI
Bookmark
Share

Authors

JZJiacui ZouYZYongchuan ZhangFLFeng Li

Discussion

Loading...

Member takes

Overview

Automated framework enhances building extraction accuracy in high-resolution urban photogrammetry, indicating a new approach.

Key Points

  • The aim is to improve the accuracy of building individualization in oblique photogrammetry 3D models.
  • Constructed a framework from orthophoto generation to dynamic 3D rendering.
  • Utilized Gated-ASPP High-Resolution Network (GA-HRNet) for enhanced segmentation.
  • Mapped extracted 2D contours back to the 3D model.
  • Achieved an Intersection over Union (IoU) of 91.25%.
  • Secured an F1-Score of 95.41%.
  • Demonstrated superior performance over models like FCN and U-Net, with IoU lead of 1.46 to 5.62 percentage points.

Cite This Study

Zou et al. (2026) studied this question.

synapsesocial.com/papers/698434f9f1d9ada3c1fb3b64https://doi.org/10.3390/app16031486
View Full Paper
Ask AI
Bookmark
Share