Related Experiment Video
Updated: Aug 14, 2026
08:27
Image Recognition and Parameter Analysis of Concrete Vibration State Based on Support Vector Machine
Published on: January 5, 2024
Vision-Based Autonomous Quadrupedal Robot for Rapid Post-Earthquake Crack-Based Building Damage Detection
Kemal Hacıefendioğlu1,2, Murat Günaydın1,2, Ayşecan Bostan1
1Department of Civil Engineering, Faculty of Engineering, Karadeniz Technical University, 61080 Trabzon, Türkiye.
Sensors (Basel, Switzerland)
|August 13, 2026
Summary
This study introduces a quadrupedal robot system for rapid earthquake damage detection in concrete structures. The automated system uses AI to identify cracks, enhancing safety and efficiency in disaster response.
Area of Science:
- Robotics and Artificial Intelligence
- Civil Engineering and Structural Health Monitoring
- Disaster Response and Management
Background:
- Post-earthquake structural assessments are critical for safety but often delayed by manual inspections.
- Engineers face significant risks when inspecting damaged, unstable buildings.
- Current methods lack the speed and automation needed for effective disaster response.
Purpose of the Study:
- To develop and evaluate a robotic system for automated, crack-based visual damage detection in reinforced concrete structures after earthquakes.
- To enhance the speed, accuracy, and safety of post-earthquake structural health assessments.
- To reduce human exposure to hazardous environments during disaster response.
Main Methods:
- A Unitree Go2 quadrupedal robot equipped with an Intel RealSense D435i RGB-D camera was utilized.
- A dataset of 3255 annotated crack images was collected from field and public sources.
- The YOLOv8n object detection model was trained and deployed on an NVIDIA Jetson AGX Xavier for real-time analysis.
Main Results:
- The robotic system achieved high detection performance in laboratory tests on reinforced concrete specimens.
- Precision, recall, and mAP@50 values for crack detection exceeded 85%.
- The system demonstrated capability for fast, accurate, and automated structural health assessments.
Conclusions:
- The proposed quadrupedal robotic system offers a viable solution for rapid, safe, and efficient post-earthquake damage detection.
- Automated visual inspection significantly reduces risks to human inspectors in hazardous post-disaster scenarios.
- Future research will focus on expanding the system's damage detection capabilities and facilitating real-world deployment.