PPE Guard Detection System

Construction sites remain among the most dangerous workplaces, where a large proportion of accidents result from head injuries, electrocution, and impact with moving or falling objects. Ensuring that workers consistently wear Personal Protective Equipment (PPE) such as helmets, vests, gloves, masks, and safety shoes is therefore essential. Traditionally, safety compliance is enforced through manual supervision or sensor-based systems, both of which have limited real-time capabilities and scalability. To overcome these limitations, this study introduces an artificial intelligence driven vision based PPE monitoring system designed for automated detection and compliance analysis. The proposed approach employs a lightweight YOLOv8 model integrated with preprocessing and postprocessing pipelines, a FastAPI based backend, and an interactive web dashboard for live visualization. The system is trained using a custom dataset and evaluated through confusion matrix analysis and precision recall metrics to ensure robustness and accuracy. Experimental trials conducted in semi-controlled conditions demonstrate the system’s ability to accurately identify multiple PPE types in real time, offering a reliable and scalable solution for improving safety management in construction environments.