An effective electricity worker identification approach based on Yolov3-Arcface

Area of Science:

• Computer Vision
• Artificial Intelligence
• Electrical Engineering

Background:

• Detection models for electric power workers in distribution rooms often suffer from low efficiency and high complexity.
• Accurate identification of personnel is crucial for safety and operational efficiency in electrical infrastructure.

Purpose of the Study:

• To develop a high-performance, lightweight electric power worker identification approach.
• To improve the efficiency and accuracy of worker detection in distribution room environments.

Main Methods:

• An improved YOLOv3 model was utilized for object detection.
• ArcFace loss function was employed as the coordinate regression loss for target boxes.
• Dimensional clustering was used to set template boxes for bounding box prediction.

Main Results:

• The improved YOLOv3 model demonstrated high performance and a lightweight architecture.
• The proposed electric power worker identification approach achieved a high-speed recognition process with accurate results.
• Experimental validation confirmed the approach's effectiveness, detection performance, and robustness.

Conclusions:

• The developed electric power worker identification method offers a significant improvement over existing models.
• The approach provides a robust and efficient solution for identifying electric power workers in distribution rooms.
• This work contributes to enhanced safety and operational monitoring in the power industry.