Research on Transmission Line Personal Protective Equipment Detection Algorithm Based on Improved YOLOv11

Area of Science:

• Computer Vision
• Edge Computing
• Artificial Intelligence

Background:

• Edge devices in transmission line inspection require real-time personal protective equipment (PPE) recognition.
• Balancing detection accuracy and computational efficiency is a critical challenge for these devices.

Purpose of the Study:

• To propose WTLS-YOLOv11n, a lightweight and interpretable detection algorithm for PPE recognition.
• To enhance real-time performance and accuracy on edge computing platforms for safety applications.

Main Methods:

• Integrated a C3K2-WTConv module using discrete wavelet transform for multi-scale feature extraction and interpretability.
• Employed a lightweight shared composite detection head for parameter reduction.
• Utilized the MPDIoU loss function to improve localization accuracy, especially for small targets.

Main Results:

• WTLS-YOLOv11n demonstrated superior detection accuracy with significantly reduced parameters and computational complexity compared to baseline models.
• The model maintained real-time inference performance on edge hardware.
• Interpretability analysis showed decisions were driven by low-frequency structural features.

Conclusions:

• The proposed WTLS-YOLOv11n offers a transparent, efficient, and trustworthy solution for automated safety supervision in electrical power operations.
• The algorithm has broader applicability to safety-critical detection tasks requiring edge deployment and interpretable decision-making.