Research On Fatigue Driving Detection Algorithm Based On Facial Feature Analysis

With people's increasing requirement of high-quality life,long-time works is needed to achieve this goal,which inevitably lead to fatigue and sleepy.A small nap is enough for people to recover in the state of static,while it is quite different when moving,for example,the sleepy drivers who has continue to drive a for a long time will probably lead to an unavoidable disaster.Fatigue can not be eliminated,but can be reminded.The occurrence of traffic accidents can be greatly avoided as long as there are certain warning at the beginning of the fatigue.In order to achieve this function,it is necessary to monitor the state of the driver in real time,and the monitoring method under non-contact state is the most appropriate way of all.In this paper,we have done some research and innovation relating to the relevant technology,besides,driver's facial features were extracted and analysis as well,the specific works are as follows:(1)By testing the four skin color segmentation methods,two kinds of them with better performance were selected for weight combinations.According to the experiments,that the combination of skin color detection can be more accuracy under the same time condition,whose pixels bias is less than half of the single skin detection.(2)we proposed a new method of driver fatigue detection,through common sense and experimental analysis,we found that the driving behavior is different under normal driving and fatigue driving,that is,the driver will be observed the rear-view mirror in a certain period of time when normally driving,while this behavior will be weakened or even forgotten when the driver is sleepy and tired.By the face localization curve,we can determine whether the driver is fatigue.(3)By study the traditional facial landmark localization algorithm,tracking the eye feature points on the basis of SDM facial landmark detection,the human eye feature point tracking based on human eye topology is proposed,which can get the position rapidly.Based on the upon theory,the detective method of driving fatigue was proposed based on the criteria of PERCLOS.The experiment was proved to have good accuracy.