Deep Learning Based Drowsiness Detection With Alert System Using Raspberry Pi Pico

Most people are losing their lives each year as a result of vehicle accidents caused by sleepy drivers. A device that can identify tiredness and notify the driver to it is necessary to prevent car accidents and save lives. Finding drowsy drivers is perhaps the most important step in stopping any traffic accident, everywhere in the globe. The goal of this work is to create alert strategy for smart vehicles which immediately halt drowsy driving problems. However, feeling tired is a normal physical phenomenon that can happen for a number of different causes. In order to stop the accident’s cause, an accurate alarm system must be developed. The system for alerting sleepy drivers that has been developed using this technique is discussed in the proposed study where Video Stream Processing (VSP) is processed using the blinking criteria based on the eyes Euclidean distance. Additionally, the face marking method is created. The results of proposed work uses Dlib, an OpenCV deep learning technique using Raspberry Pi with a mounted camera for accurate detection, showing good results for detecting drowsiness, reducing the accident occurrence on the roads. The facial recognition algorithm developed by Haar starts with collected photographs as its input and outputs faces that are identified.