Detecting Driver Drowsiness State Based On Video Image

Recently years, traffic accidents distributed frequently have brought to people’slives great pain and injury, Governance of the state of traffic safety hazard has becomethe drawbacks of the current government safety regulators. In addition to the growingsafety awareness, to develop consciously abide by the traffic safety and order, butalso need to rely on the power of science and technology, when people is not enoughin the case of self-control, with external devices warning to remind each traffic driversafe driving. In many factors that lead to accidents, driver fatigue driving is one of themost important aspects. Transport Research Laboratory supposes, Driver fatigue causeof road traffic accidents accounted for10%of all traffic accidents. According tostatistics, Worldwide each year up to60million deaths due to traffic accidents causedand the direct economic losses of about$12.5billion,57%of these catastrophicaccidents related to driver fatigue driving. Research driver fatigue warning has a veryimportant social significance, and has a mass of important guiding significance forpeople’s daily life; this is also article topics and research purpose.This paper presents a method for real-time monitoring of driver fatigue warningin potential traffic accidents by face multi-feature method. Substituting alternative tothe status of the previous studies relying solely on the driver’s eye state or the driver’sface to judge and prediction, The article uses information fusion method combiningthe driver’s eyes with mouth state to judge fatigue, When the driver ware the eyesdriving or traveling in the dark, the paper detection method can play a good earlywarning. AdaBoost algorithm is used to detect face region due to its high correct rateand good robustness. According to the structural characteristics of the face and theexperience knowledge of the human, extract the position of the eyes and mouth blockimage from the face image, to detect the eyes state and mouth state, developing a newdriver’s state of motion rules, the paper takes new fatigue detection method PATECP(Percentage And Time that Eyelids Cover the Pupils) and PATMIO (Percentage AndTime that Mouth Is Open) to verify the driver’s state. The tests with actual drivingvideo shows that our approach based on information fusion method of eyes state andmouth features makes the conditions of recognizing the driver’s drowsy state wideraccurate and the driver fatigue detection can adapt to various conditions.In summary, this paper do a preliminary study of driver fatigue warningintegrating of multi-feature detection of the face and the mouth of the driver, Andconstruct a real-time monitoring and warning system for the target, Real-time monitoring of the state of motion of the drivers. And the algorithms involved in thepaper have been realized by program which presents strong stability and satisfactoryresults in the process of many sample tests.