Driveshield

Abstract

The AI Dashcam Project is an innovative system designed to enhance road safety by integrating advanced artificial intelligence and real-time video analytics into a vehicle monitoring platform. Utilizing a Raspberry Pi, the system incorporates key features such as access control, drowsiness detection, collision detection, and an alert system for real-time monitoring and reporting. Access Control is implemented using the YOLO (You Only Look Once) algorithm to provide real-time facial recognition, distinguishing between registered drivers and unauthorized individ- uals attempting vehicle access. Drowsiness Detection employs OpenCV and pre-trained Convo- lutional Neural Networks (CNNs) to monitor the driver�s facial expressions and detect signs of fatigue, including yawning and prolonged eye closure. This system helps to mitigate risks by is- suing timely alerts before dangerous situations occur. Collision Detection uses continuous video recording alongside the Single Shot MultiBox Detector (SSD) algorithm to detect collision events, storing buffered video of incidents to capture the moments before and after an accident for forensic purposes. Additionally, an Alert System powered by Wi-fi communication ensures that critical notifica- tions regarding driver drowsiness or collision events are transmitted to a mobile app for remote monitoring, even in areas with limited internet connectivity. This feature allows fleet managers or authorized personnel to respond quickly and effectively. This AI dashcam system aims to provide a cost-effective and reliable solution for increasing road safety through intelligent vehicle monitoring and real-time interventions, with the Raspberry Pi acting as the central processing hub for the camera and AI-based analytics.