Detection method of PPE wearing and small target positioning for offshore operators: The improved YOLOv11 model and targets recognition
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.This study enhances offshore operator safety by improving Personal Protective Equipment (PPE) detection using an advanced YOLOv11 model. The new method significantly boosts accuracy for small items like earplugs and gloves, reducing missed detections.
Area Of Science
- Computer Vision
- Artificial Intelligence
- Occupational Safety
Background
- Offshore environments pose unique challenges for monitoring Personal Protective Equipment (PPE) usage.
- Existing methods struggle with insufficient data, incomplete categories, and low accuracy, especially for small PPE items.
- Accurate PPE detection is critical for operator safety in high-risk offshore settings.
Purpose Of The Study
- To develop an improved deep learning model for accurate PPE detection among offshore operators.
- To address the specific challenges of detecting small and easily missed PPE items like earplugs and gloves.
- To enhance the overall reliability of PPE wearing condition monitoring systems.
Main Methods
- An improved YOLOv11 model was developed, integrating Spatial Pyramid Pooling-Fast (SFEAF) and Token Statistics Self-Attention (TSSA) modules.
- A Normalized Wasserstein Distance (NWD) loss function was incorporated to refine detection capabilities.
- A cascaded network combining YOLOv11 and YOLOv11-Pose was proposed for enhanced small target detection, utilizing key point extraction and region refinement.
Main Results
- The optimized YOLOv11 model achieved a 1.8 percentage point increase in mAP@0.5 for all targets.
- Precision rates for small targets (earplugs, gloves) improved by up to 5.2%.
- The cascaded network demonstrated an average Missed Detection Recovery Rate (MRR) of 56.64% for small targets.
Conclusions
- The enhanced YOLOv11 model significantly improves PPE detection accuracy and reliability for offshore operators.
- The cascaded detection approach effectively addresses the challenge of missed detections for small PPE items.
- This research provides a robust solution for improving safety compliance in demanding offshore work environments.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
