Optimized Real Time Drowsy Driver Warning System Using Raspberry Pi

Abstract

Accidents are undesirable, incidental and unexpected events that can be prevented if they are recognized and acted upon, prior to their occurrence. Driver drowsiness is one of the major causes of road accidents. A solution to this problem is inclusion of a drowsiness warning system in vehicles to warn the driver of drowsiness. The main objective of this paper is to develop a non-intrusive real-time drowsiness warning system by monitoring eye state of the driver. The proposed system uses eye-aspect-ratio and is implemented in Raspberry Pi. This system is optimized using three techniques, by re-sizing the frame to be processed, using a single eye for drowsy detection and by not processing every frame. The selection of frames is based on analyzing the redundancy nature of video frames. This system is designed by considering several factors to minimize frame processing latency to match with Raspberry Pi’s processing speed.