ESE-YOLOv8: A Novel Object Detection Algorithm for Safety Belt Detection during Working at Heights

Area of Science:

• Computer Vision
• Artificial Intelligence
• Occupational Safety

Background:

• Manual supervision of workers at heights is challenging and prone to human error.
• Ensuring compliance with safety belt usage is critical for preventing accidents.

Purpose of the Study:

• To develop an automated object detection system for monitoring safety belt usage.
• To introduce and evaluate a novel object detection model, ESE-YOLOv8, for enhanced worker supervision.

Main Methods:

• Developed ESE-YOLOv8, integrating Efficient Multi-Scale Attention (EMA) and GSConv.
• Employed Efficient Intersection over Union (EIoU) loss function for improved localization.
• Conducted experiments and ablation studies to validate model performance.

Main Results:

• ESE-YOLOv8 achieved 92.7% average precision at IoU 50% and 75.7% at IoU 50%-95%.
• Outperformed baseline YOLOv5 and demonstrated competitive results against state-of-the-art models.
• Ablation experiments confirmed the effectiveness of proposed enhancements.

Conclusions:

• ESE-YOLOv8 offers a robust and accurate solution for automated safety belt supervision.
• The model significantly improves upon existing methods, enhancing worker safety at heights.
• This technology has the potential to revolutionize safety monitoring in high-risk work environments.