A Real-Time IoT-based Model to Detect and Alert Security Guards’ Drowsiness

Abstract

In the security industry, it is critical to ensure that security guards remain alert and attentive throughout their shifts. Drowsiness and inattention can lead to security breaches and endanger the safety of the premises and the people within them. To address this issue, a hybrid system is being developed to detect security guards' drowsiness and alert them using sound buzzers and water sprinklers to prevent security breaches. The system uses advanced machine learning and deep learning techniques like OpenCV and DCNN, along with a UHD camera, to detect signs of drowsiness using algorithms like Eye Aspect Ratio/Mouth Aspect Ratio (EAR)/(MAR). By analysing behavioural indicators, the system determines whether a security guard displays signs of drowsiness and alerts them using sound buzzers to remain attentive. Overall, the hybrid system provides an effective solution to enhance security guard monitoring and prevent potential security threats caused by drowsiness and inattention.