Study On The Application Of Fatigue Driving Warning System

The rapid development of the economy has promoted the sharp increase in the number of cars,the car brings great convenience to people’s daily life.A series of traffic problems have happened because of the human factors.One of the main factors of traffic accidents is the driver’s fatigue driving.Therefore,it is of great significance to study a fatigue warning device that can be used in automobiles to reduce the occurrence of traffic accidents.At present,There is no one solution that can fully meet the demand of the market.The application of the fatigue driving warning system in the market is not universal,mainly because it is difficult to achieve a balance between the recognition accuracy and the low cost.Based on the above considerations,this project has designed a high-recognition,low-cost fatigue driving warning system program and implemented it using engineering methods.First of all,we choosed the AdaBoost algorithm which was combined with Haar feature that is most suitable for the practical application to perform face localization,the combination of Camshift and AdaBoost algorithm was also employed to improve real-time nature of face positioning.Then the AdaBoost algorithm based on Haar feature for the eye location in the semi face region was adopted.In addition,the state of the human eye was recognized through the projection method,which can’t completely identify the closed eye.This paper combined the ellipse fitting and projection method to improve the accuracy of closed eye state recognition.The PERCLOS algorithm was used to judge the fatigue procedure of human eyes.Secondly,the FPGA hardware acquisition card circuit was designed.The hardware interconnection between AM5728 and Camera Link camera,which provides a guarantee for AM5728 to video capture of HD Camera Link camera,was realized through the circuit’s schematic design,circuit board making and debugging.This paper also studied the video acquisition system of the USB camera based on UAC protocol,which is convenient for applying the fatigue driving warning system to more occasions.Finally,we transplanted the image algorithm of fatigue driving warning system to the AM5728 platform,and optimized the real-time nature and accuracy of the algorithm.In general,this paper provided a set of high accuracy and low cost research methods for fatigue driving warning system.