Font Size: a A A

The Detection Of Driving Fatigue Based On Facial Features

Posted on:2014-01-04Degree:MasterType:Thesis
Country:ChinaCandidate:L W ZhangFull Text:PDF
GTID:2268330401488913Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
One of the main reasons for traffic safety is driver fatigue.In order to preventtraffic accidents and protect security effectively，it is necessary and meaningful todetect driver fatigue.This paper proposes the detection of driving fatigue based onthe characteristics of lips and faces.The main idea is as follows:1.The detection and prediction of human faces.The gaussion model is takenadvantage of to detect the face.The calman filter is exploited to predict the locationof the face.Experimental results show that the effect and the speed of facepositioning can be improved.2.Eye positioning and the judgement of a quick doze.The region growing andmorphological operations are combined to establish the eye position.Thepercentage of the black pixels’ number in the eyes contour is used to determinewhether eyes are closed.Then，the corresponding PERCLOS is calculated to detectthe driver’s dozing off.Experimental results show that the effect of dugement ofdozing of by this method is good.3.The candidate lips areas are located according to the best threshold of colorspace and the facial binary image.Then，we detect whether the mouth is opened andhow long the mouth keep openging.so we can judge whether the driver isyawning.Experimental results show that the effect of dugement by this method isgood.Finally，driving fatigue is decided based on two facial features.Experimentalresults show that the effect of judgement based on two facial features is better.
Keywords/Search Tags:Driving fatigue, Face detection, Lip detection, Skin color model, Kalman prediction
PDF Full Text Request
Related items