Design And Implementation Of Driver Fatigue Detection System Based On Multi-features

With the continuous development of our country’s economy, the number of motorvehicles is increasing rapidly,as a result,the traffic accidents also increaseaccordingly.According to the relevant statistics from some research institutions,we knowthat driver’s fatigue is one of the important cause of traffic accidents.Therefore developeda real-time and accurate system to detect the driver’s fatigue state has important practicalsignificance and social influence.After the analysis of the research on driver fatigue detection both at home andabroad,we use the digital image processing technology as the foudation,combined withrelevant knowledge in the field of artificial intelligence,machine learning,machinevision,and then we propose the fatigue detection system based on multiple features.Thefatigue detection system bases on the success of the face detection,and it achieves thepositioning and detection of the sunglasses,the positioning of the eyes and the extractionof eye’s fatigue paramaeters,the positioning of the mouth and the extraction of mouth’sfatigue paramaeters.In the aspect of the eye’s positioning,we use the human eye detection algorithmwhich based on AdaBoost algorithm and eye tracking algorithm which based on the headmovement in combination to pinpoint the eye area.And then we propose PERCLOSparameter,ECT parameter and EBF parameter,which could stand for the eye’s fatiguestate.In the aspect of the sunglasses’s positioning,we use the detection algorithm whichbased on the location characteristics of sunglasses,according to the position of people wearsunglasses in daily life to accurately pinpoint the sunglasses area,and then the systemdetermines whether the people is wearing a sunglasses.In the aspect of the mouth’spositioning,we use the human mouth detection algorithm which based on AdaBoostalgorithm and the detection algorithm which based on the mouth position feature incombination to pinpoint the mouth area. And then we propose PMRCLOS parameter andYawnFreq parameter,which could stand for the mouth’s fatigue state.Finally we achieve the fatigue detection system based on multiple features,and wetest the system.From the test result,we can find that the system implements the functional requirements of fatigue detection system very well,and the system has high accuracy andreal-time performance.So we could know that the fatigue detection system based onmultiple features has high scientific and effective,it is useful for the detection of thedriver’s fatigue state.