Related Experiment Video
Updated: Jun 9, 2025
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
10.6K
Survivor detection approach for post earthquake search and rescue missions based on deep learning inspired algorithms
Rajendrasinh Jadeja1, Tapankumar Trivedi2, Jaymit Surve2
1Department of Electrical Engineering, Marwadi University, Rajkot, 360003, Gujarat, India. rajendrasinh.jadeja@marwadieducation.edu.in.
Scientific Reports
|October 24, 2024
Summary
This study introduces a snake robot with deep learning for detecting earthquake survivors under debris. The YOLOv10 algorithm showed high accuracy (98.5%) and speed (15 ms), improving search and rescue operations.
Area of Science:
- Robotics
- Artificial Intelligence
- Disaster Management
Background:
- Effective search and rescue (SAR) is vital after earthquakes.
- Survivor detection under debris presents significant challenges due to occlusion and clutter.
Purpose of the Study:
- To develop and evaluate a novel survivor detection system using a snake robot and deep learning.
- To compare the performance of Faster R-CNN, SSD, and YOLO algorithms for detecting trapped humans.
Main Methods:
- A snake robot equipped with deep learning object identification algorithms was utilized.
- A new dataset of 200 images depicting trapped survivors in cluttered environments was created.
- Performance was evaluated based on detection accuracy, confidence intervals, and running time.
Main Results:
- The YOLOv10 algorithm achieved a mean average precision (mAP) of 98.4 at an IoU threshold of 0.5.
- YOLOv10 demonstrated an accuracy of 98.5% with an inference time of 15 ms.
- Algorithms were validated on images of survivors under various occlusion conditions.
Conclusions:
- The YOLOv10 algorithm offers a highly accurate and efficient solution for survivor detection in post-earthquake SAR.
- The developed snake robot system shows promise for enhancing the speed and effectiveness of rescue operations.
