Design And Implementation Of Driver Fatigue Detection System Based On DSP

With the development of social economy,the traffic safety problem is becoming more and more prominent.Through the research on a large number of traffic accidents,it is found that driver fatigue driving is one of the high threat factors of driver traffic accident.Due to individual difference,environment influence and real time,the existing fatigue detection products have not been completely popularized,so it is of great practical significance to develop an effective driver fatigue detection system.This paper analyzes and compares three kinds of driver fatigue detection methods,which are popular at present,combining with the environment of the laboratory,by studying the relevant documents and materials of fatigue driving detection system at home and abroad.In this paper,a simple and feasible method based on DSP video image processing is selected to detect driver fatigue.The main work is as follows:The realization scheme and framework of the detection system are determined,and the image preprocessing methods,including image enhancement algorithm and binary image processing,are studied in the realization of the system.Face detection is realized by fully utilizing face characteristics and combining with AdaBoost Haar algorithm to realize face detection.After accurately locating the face,the eyes are detected step by using the idea of coarse and fine double detection,and the position of human eyes is located with coarse step by step using gray scale integral projection method.Using the Sobel edge detection operator and the improved Hough algorithm to locate the human eye accurately,then extracting the eye features and calculating the fatigue parameters,such as blink frequency,etc.The fatigue state of driver is judged and warned by P80 criterion which occupies a certain proportion of eye closure in unit time.In this paper,the method of driver fatigue detection based on DSP video image processing is a better solution to the driver's personal safety problems caused by fatigue driving.The hardware platform of TI C6000 series TMS320DM642 DSP processor is adopted,and a function knot is carried out in the CCS environment.The structure,C programming style and compilation are optimized to improve the running speed of the algorithm and realize real time fatigue detection,and the response time of fatigue detection can not exceed 38 ms.Compared with the traditional method based on PC,the implementation method is small,practical and highly efficient.It does not affect the driver's normal driving,reduces the cost of the system and ensures the real-time performance of the system.