The construction industry faces significant challenges, including a high prevalence of occupational incidents, often involving fires, explosions, and burn-related accidents due to worker non-compliance with safety protocols. Adherence to safety guidelines and proper utilization of safety equipment are critical to preventing such incidents and safeguarding workers in hazardous work environments. Consequently, a monitoring system tailored for construction safety during welding operations becomes imperative to mitigate the risk of fire accidents. This paper conducts a brief analysis of OSHA rules pertaining to welding work and introduces the iSafe Welding system, an advanced real-time safety monitoring and compliance enforcement solution designed specifically for construction site welding operations. Harnessing the real-time object detection algorithm YOLOv7 in conjunction with rule-based scene classification, the system excels in identifying potential safety violations. Rigorous evaluation, encompassing precision, recall, mean Average Precision (mAP), accuracy, and the F1-Score, sheds light on its strengths and areas for improvement. The system showcases robust performance in rule-based scene classification, achieving high accuracy, precision, and recall rates. Notably, the iSafe Welding system demonstrates a formidable potential for enhancing construction site safety and regulatory compliance. Ongoing enhancements, including dataset expansion and model refinement, underscore its commitment to real-world deployment and its strength in ensuring worker safety
Chung Ang University, Korea (the Republic of) - ORCID: 0000-0003-2257-290X
Chung Ang University, Korea (the Republic of) - ORCID: 0000-0002-6909-5189
Chung Ang University, Korea (the Republic of)
Chung Ang University, Korea (the Republic of) - ORCID: 0000-0002-8192-340X
Chung Ang University, Korea (the Republic of) - ORCID: 0000-0002-6559-8782
Chung Ang University, Korea (the Republic of) - ORCID: 0000-0003-2256-300X
Chapter Title
iSafe Welding System: Computer Vision-Based Monitoring System for Safe Welding Work
Authors
Syed Farhan Alam Zaidi, Rahat Hussain, Muhammad Sibtain Abbas, Jaehun Yang, Doyeop Lee, Chansik Park
DOI
10.36253/979-12-215-0289-3.66
Peer Reviewed
Publication Year
2023
Copyright Information
© 2023 Author(s)
Content License
Metadata License
Book Title
CONVR 2023 - Proceedings of the 23rd International Conference on Construction Applications of Virtual Reality
Book Subtitle
Managing the Digital Transformation of Construction Industry
Editors
Pietro Capone, Vito Getuli, Farzad Pour Rahimian, Nashwan Dawood, Alessandro Bruttini, Tommaso Sorbi
Peer Reviewed
Publication Year
2023
Copyright Information
© 2023 Author(s)
Content License
Metadata License
Publisher Name
Firenze University Press
DOI
10.36253/979-12-215-0289-3
eISBN (pdf)
979-12-215-0289-3
eISBN (xml)
979-12-215-0257-2
Series Title
Proceedings e report
Series ISSN
2704-601X
Series E-ISSN
2704-5846