samveg12 / drive-safe

Drive-Safe

We propose a system to monitor driver alertness and awareness using smartphone cameras. The system matches the face of the driver to the photo on his/her driver’s license. It detects the drowsiness of the driver using a real-time video stream. It also classifies driver actions into 9 commonly occurring distraction categories or as safe driving. The front facing video stream of the driver’s phone is also used for monitoring road safety behaviour by estimating the distance to surrounding vehicles on the road and other risks such as pedestrians and signal violations. We have also developed a model that can classify the roads into one of the 4 categories - Good, Poor, Satisfactory and Very Poor using CNN. Lastly, the system alerts drivers to the possibility of increased accident risk based on various factors such as location, time of day, etc.