Related Experiment Video
Updated: Jun 21, 2025
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
11.1K
Surface Defect-Extended BIM Generation Leveraging UAV Images and Deep Learning
Lei Yang1,2, Keju Liu3, Ruisi Ou3
1Key Laboratory of Urban Land Resources Monitoring and Simulation, Ministry of Natural Resources, Shenzhen 518034, China.
Sensors (Basel, Switzerland)
|July 13, 2024
Summary
This study introduces an accurate deep learning method for detecting building defects from drone images. It maps these defects onto Building Information Models (BIM) for improved digitalization in construction.
Area of Science:
- Construction Science
- Computer Vision
- Digitalization in Construction
Background:
- Building defect inspection is crucial for construction digitalization.
- Drone and AI technologies offer advanced inspection tools.
- Integrating UAV defect data into Building Information Modeling (BIM) faces accuracy and coordinate challenges.
Purpose of the Study:
- To develop an accurate defect detection method for UAV images.
- To establish a coordinate mapping technique for integrating defect data into BIM.
- To create an enriched BIM model with surface defect information.
Main Methods:
- A deep learning approach combined with transfer learning for defect detection.
- A texture mapping method to translate image coordinates to BIM project coordinates.
- Projection of detected defects onto BIM surfaces to create a Surface Defect-Extended BIM (SDE-BIM).
Main Results:
- High accuracy in defect detection using the deep learning method.
- Successful mapping of defects from UAV images to the BIM model.
- Demonstrated applicability in a real-world case study at Nantong University.
Conclusions:
- The proposed method effectively integrates UAV-based defect inspection data into BIM.
- This approach enhances the digitalization of existing buildings.
- The validated methods show broad applicability for diverse building inspection tasks.
