Related Experiment Video
Updated: Nov 7, 2025
11:34
Subsurface Defect Localization by Structured Heating Using Laser Projected Photothermal Thermography
Published on: May 15, 2017
11.3K
Deep Learning-Based Concrete Surface Damage Monitoring Method Using Structured Lights and Depth Camera
Hyuntae Bang1, Jiyoung Min2, Haemin Jeon1
1Department of Civil and Environmental Engineering, Hanbat National University, Dongseodae-ro 125, Daejeon 34158, Korea.
Sensors (Basel, Switzerland)
|April 30, 2021
Summary
This study introduces an automated system for detecting concrete structural damage like cracks and delamination using deep learning and a depth camera. The method achieves high accuracy in identifying and quantifying surface defects, crucial for infrastructure health monitoring.
Area of Science:
- Civil Engineering
- Computer Vision
- Artificial Intelligence
Background:
- Aging infrastructure necessitates automated structural health monitoring.
- Surface damage (cracks, delamination, rebar exposure) is key to assessing concrete structure condition.
- Declining construction workforce drives demand for automated solutions.
Purpose of the Study:
- To propose a deep learning-based system for detecting and quantifying surface damage on concrete structures.
- To utilize structured lights and a depth camera for precise damage assessment.
- To automate the monitoring of structural integrity.
Main Methods:
- A monitoring system combining four lasers and a depth camera was developed.
- Image homography was used for calibrating images when the structure and sensor are not parallel.
- Faster RCNN with Inception Resnet v2 architecture was employed for damage detection (cracks, delamination, rebar exposure).
- Damage quantification involved analyzing laser beam positions and measured distances.
Main Results:
- The system successfully detected three types of surface damage.
- Structural damage was identified with a high F1 score of 0.83.
- Quantified relative error for damage assessment was below 5% median value.
Conclusions:
- The proposed deep learning approach effectively automates structural damage monitoring.
- The system demonstrates high accuracy and reliability in detecting and quantifying concrete surface defects.
- This technology offers a promising solution for maintaining aging infrastructure.
