Object Detection Technologies in Power Construction Safety Monitoring: Challenges, Applications and Future Directions

The shift toward high-quality power construction intensifies safety challenges due to complex environments, dense personnel, and high-risk operations. Traditional manual monitoring lacks efficiency and real-time capabilities. This paper reviews advancements in object detection technologies for power construction safety monitoring, analyzing key challenges: multi-target detection, small-target recognition, and dynamic occlusion. We compare two-stage (region proposal-based) and one-stage (end-to-end) frameworks in scenarios like safety gear detection and high-altitude violation analysis, evaluating their accuracy-speed-robustness trade-offs. Integration with pose estimation and behavior analysis enhances complex violation recognition. Future directions emphasize multimodal fusion, lightweight deployment, and adaptive optimization for intelligent monitoring systems.