Heterogeneous Multi-Model Ensemble for PPE Detection in Construction Environments

The construction industry remains one of the most hazardous work environments, with a fatality rate of 25.6 per 100,000 workers, significantly exceeding the average across all industries. PPE compliance is crucial for worker safety, yet monitoring adherence remains challenging in dynamic construction environments. This paper presents an automated PPE detection system utilizing an ensemble deep learning models to enhance workplace safety monitoring. Our approach combines three advanced architectures with Yolov11, RTDETRv2, and HyperYolo. The individual model predictions are integrated using WBF to improve detection robustness. We evaluate our system on a comprehensive dataset of 4,135 professionally annotated images encompassing critical PPE categories including hard hats, safety vests, protective gloves, and safety boots, along with their corresponding absence classes. The proposed ensemble achieves superior performance with a precision of 0.765, recall of 0.735, mAP@50 of 0.760, and mAP@50:95 of 0.440, outperforming individual models across all evaluation metrics. The results demonstrate the effectiveness of multi-model fusion for automated PPE detection. This research contributes to the advancement of intelligent safety systems that can significantly reduce workplace injuries and fatalities through automated PPE compliance verification.