Motor Vehicle Driver Video Image Fatigue Detection Algorithms

In order to test the fatigue state of motor vehicle drivers in the driving process as well as assist the regulatory supervision control of the transport sector, avoiding traffic generation,This paper has studied one fatigue simulation test system of the drivers’ face recognition and detection of the human eye through computer simulation.The system combines with VC++6.0and OpenCV1.0as the programming development platform, which has completed the program and the human eye driver fatigue detection function through the preparation of live video capture program and face detection program.Firstly, the fatigue test status quo of domestic and foreign is outlined in the paper, and then the main thesis work and the overall organizational structure are described.By using MFC on dialog-based video image programming,the live video acquisition and extraction of a single frame in building VC++6.0and OpenCV1.0has been realized.Through the analysis of commonly used detection algorithms at present, the classification learning algorithm is choosed to realize the face detection and the positive and negative face image normalization are used as samples to train face images and adjust the proportion of the number of positive and negative samples according to the need and get the face detection and classification (.xml) file, By the method of loading this classifier file, faces are detected in a single frame. Based on the face recognition, the eyes experience characteristics and geometric distribution as well as the golden eye position is applied to realize the eye location simply and ingeniously, which is prepared for eye position judgment. The idea of eye rotation speed is proposed to judge the degree of driver fatigue and achieve real-time monitoring of driver fatigue.These related algorithms,the implementation process and results,the subject code above are presented in the methods narrated.Finally,through debugging and operation of the code of the system design, some experimental results are given,including the acquisition of scene video, face recognition and eye detection etc.The results show that:the algorithm can meet the design requirements with fast speed, high accuracy and strong adaptability, which can satisfy the request of real-time application.The program is stable and has a certain practicality,which can complete the detection task of driver fatigue.