Real-time driving monitoring system for new drivers using 360° camera with raspberry PI

Cho, Wei Tuck (2020) Real-time driving monitoring system for new drivers using 360° camera with raspberry PI. Final Year Project, UTAR.

Preview

PDF Download (8Mb) | Preview

Abstract

The Advanced Driver Assistance System (ADAS) is invented and installed to the vehicles to protect drivers and reduce car accident by monitoring the road condition. However, not all of the vehicles can install ADAS system due to the system complexity and the great price of the required sensors. Vision-based ADAS is affordable for most of the driver as it needed only cameras as sensors to monitor the car surrounding. Car accidents happened frequently among the teenagers. This is because they were not familiar with the road condition and lack of driving skills after they newly get their driving license. Some blind spots might difficult to be noticed by the new drivers during the driving process. Drivers might hit other vehicles and lead to small collision or serious tragic accident. Therefore, a driving monitoring system is needed to keep an eye on the road condition and provides guidance to the drivers. The proposed system provides a surround view of the vehicle and able to alert the driver when the vehicle is near to other vehicles. Raspberry Pi is the main component of the proposed system and 3 wide angle Pi camera are used to capture the surround view. Besides, Forward Collision Warning (FCW) feature is applied to this system by using a pre-trained TensorFlow model for vehicle detection and an approximate distance algorithm to determine whether or not the distances between vehicles are closer enough to trigger the alert action.

Actions (login required)