Design And Implementation Of Driver’s Fatigue Detection System Applications Based On Embedded System

With the rapid development of national economy in recent years，the city as wellas high-speed road motor vehicle ownership is increasing dramatically.In theworld,every year due to the driver fatigue which leads to traffic accidents andcasualties of tens of thousands.In our country,fatigue driving which is defined as oneof the three main reasons of road traffic accidents by the traffic controldepartment,brings many bad effects to people’s life and property.Therefore,how toprevent driver fatigue effectively to reduce traffic accidents caused by fatiguedriving,has very important practical significance.In this paper,USB camera is used to get driver’s real-time facial videoimage,analysis of the states of eyes and mouth to determine whether the driver isfatigue or not.In this paper,i design and implement a driver fatigue detection based onembedded system with high accuracy and stability.Firstly,the fatigue detection systemuses algorithms are given in detail. The system uses the core algorithm is based onAdaboost cascade classifier algorithm Haar features, to achieve video face detectionand eye detection by the algorithm. Then the divided face image, the image is ahorizontal projection and vertical projection, located the position of the nozzleportion. Finally, two PERCLOS and yawning fatigue evaluation criteria, to achievedriver fatigue state judge. Through trial and error and parameter adjustment, on thePC side has a higher detection rate of fatigue fatigue detection system.Algorithm design, the realization of the transplant program at the ARM platform.In this paper, Tiny4412development boards and ordinary USB camera as thehardware platform, the first port of the Linux operating system, Qt graphical userinterface library, and computer vision library OpenCV on the platform. The trial thenget a higher fatigue detection efficiency of the simulation program in the PC side,transplanted to Tiny4412development board. Upon completion of porting to ARMdevelopment board as experimental platform, the validity of the algorithm andreal-time analysis. Experiments show that the system is designed to achieve the desired goal.