Solution Idea Explanation - reyanshsolis/safety_driving_assist GitHub Wiki

The Alertness level of drivers can be estimated with the use of computer vision based methods. The level of fatigue can be found from the value of PERCLOS (It is the ratio of closed eye frames to the total frames processed). In this work we have developed a real-time system which is able to process the video onboard and to alarm the driver in case the driver is not alert and even assist Braking during difficult situations. Using pre-trained model which detects face and give landmarks of different features, distance between lips and Eyes Aspect Ratio we study the alertness of the Driver. During night time active Near Infrared (NIR) illumination will be used.

Autonomous Braking Algorithm used is a very effective way of bringing the vehicle to a stop immediately, in situations of obstacle appearing suddenly in front of the car, and minimising the damage.

As road accidents occurs within a very short time frame, thus a delay of milliseconds can be a factor of life and death, So this application requires fast processing and accelerated functions. This is where the role of PYNQ comes. The Idea of using PYNQ is to accelerate the heavy computation part of Image Processing by offloading it to FPGA.