Driver Fatigue Detection

Description

Many people working in different kinds of professions require undisturbed concentration for prolonged periods of time. It is therefore, vital for drivers to maintain an eye on the road at all times. By ensuring this, one can react to any events that occur suddenly. In the recent years, the reason for maximum number of vehicle accidents in the world is driver fatigue. Hence, it is important to develop systems that have the ability to reduce the number of car accidents that are caused due to fatigue by detecting any psychological or physical condition changes in the driver and notifying the driver about it. But the process of developing that kind of machines faces lots of difficulties that pertain to speedy and apt recognition of the symptoms for driver’s fatigue. Driver fatigue can be directly measured in a way by analyzing the current condition of the driver that is fatigue. Our project aims to develop a model for detection of fatigue. This driver fatigue detection system helps in identifying the fatigue and issuing a timely alarm to alert the driver. This real time system captures incessant pictures from a streaming video and measures what is the state of the eye based on an algorithm and notifies a warning message when necessary. Unlike the various existing approaches for identifying the fatigue of an individual, our approach proves to be completely non-intrusive by affecting the driver in any way and as a result providing us with the exact state of the driver very precisely. In order to detect the fatigue, the PERCLOS(per closure value of eye) is considered. When the PERCLOS value outruns a given value, the driver is pointed out to be sleepy. To implement this system Haar-cascade, several OpenCv libraries are also used.