Related Experiment Video
Updated: Jul 19, 2025
09:17
Surrogate Model Development for Digital Experiments in Welding
Published on: March 28, 2025
1.0K
Welding Spark Detection on Construction Sites Using Contour Detection with Automatic Parameter Tuning and
Xi Jin1, Changbum Ryan Ahn1, Jinwoo Kim2
1Department of Architecture and Architectural Engineering, Seoul National University, Seoul 08826, Republic of Korea.
Sensors (Basel, Switzerland)
|August 12, 2023
Summary
This study introduces a new computer vision method for real-time welding spark detection to prevent construction fires. The deep learning model accurately identifies sparks, enhancing fire safety surveillance.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Fire Safety Engineering
Background:
- Welding sparks are a major cause of construction site fires.
- Existing fire detection systems lack real-time tracking of small, hazardous sparks.
- Computer vision offers potential for enhanced construction site fire monitoring.
Purpose of the Study:
- To develop a novel real-time welding spark detection method using computer vision and deep learning.
- To improve the accuracy and reliability of fire detection systems for construction environments.
- To address the challenge of early detection of fire precursors like welding sparks.
Main Methods:
- Developed a real-time contour detection method with deep learning parameter tuning.
- Utilized a convolutional neural network for automatic hue saturation value optimization.
- Implemented novel filtering techniques, including non-welding zone and contour area-based filters.
Main Results:
- Achieved 74.45% precision and 63.50% recall on welding spark images after noise removal.
- Demonstrated 95.2% accuracy in detecting the peak number of welding sparks in video datasets.
- The model effectively filters out noise like flashing and reflection light.
Conclusions:
- The proposed automated welding spark detection method shows significant potential for enhancing fire surveillance.
- This technology can contribute to reducing fire incidents at construction sites.
- Further development can lead to more robust and reliable fire prevention systems.
