| At this stage,with the per capita car ownership continues to rise,traffic accidents have become increasingly frequent.At the national and global level,fatigue driving is one of the main causes of traffic accidents.If the state of the driver can be detected in real time,abnormal driving state in time with the sound or light signals in a timely manner to remind the driver,the traffic safety can be effectively improved.This paper aims to design a complete set of driver fatigue detection system,including the hardware platform of the entire system and software algorithms based on the hardware platform.This paper introduces the research background and significance of driver fatigue detection system,and summarizes the research status of related technologies at home and abroad through a lot of literature reading,and introduces several domestic and foreign mature products in detail.Then,the relevant technical principles of fatigue driving detection are expounded.The fatigue driving detection technology based on facial expression features is studied,including the technology of face detection,eye detection and mouth detection,and the criterion of fatigue state is given.Then,the hardware platform of the system is designed.The modular design idea is used,including the power supply system,DSP data processing module,external memory expansion module,video acquisition module,video output module and alarm module.The interface driver is designed for each module.The system cmd file and the secondary Boot Loader program are designed.Then,the software program of the system is designed,and the AdaBoost algorithm based on Haar feature is explained.Based on this algorithm,the cascade classifiers of face,eyes and mouth are trained.The three facial features are detected and positioned by trained cascade classifiers.According to P80 and R3 standard to calculate the degree of closure of the eye and opening of mouth,the driver fatigue state is determined by the value of PERCLOS and PERYAWN.Finally,the hardware and software modules are debugged respectively,and the joint debugging of the hardware and software is carried out.According to the hardware and software debugging of the driver fatigue detection system,the hardware platform of the system has the characteristics of normal function,strong processing power and fast running speed.The algorithm has the advantages of fast speed,high accuracy,and can process 16 pictures per second,and the recognition rate of fatigue state is more than 90%,which verify that the algorithm is fast and effective. |