Research On Fatigue Driving Detection Method Based On Multi Feature Factors

With the improvement of people’s economic level,the number of vehicles on the road is increasing.At the same time,the incidence of traffic accidents is also increasing gradually.According to the survey,one of the main causes of traffic accidents is fatigue driving.Drivers will be tired due to driving for a long time.Mild fatigue will narrow the driver’s vision,Miss road information,seriously lose driving ability and threaten the driver’s life and safety.In order to remind drivers to drive correctly and reduce traffic accidents,an efficient and accurate fatigue driving early warning system is essential for safe travel.This paper mainly studies the method of fatigue driving detection based on multi feature factors.The camera is used to collect the driver’s state information in real time,and the computer vision technology is used to analyze the face features and head posture.The driver’s fatigue state is judged in combination with the driver’s physiological characteristics such as blinking,yawning,dozing and nodding.The main research contents of this paper are as follows:(1)The fatigue detection system mainly uses computer vision technology to collect the fatigue characteristics of the driver’s face and head.Firstly,the video frame is extracted from the camera or local video through opencv,the image of each frame is extracted from the cyclic video frame,the dimension of the image is expanded,the image is grayed,and the face recognition and detection are carried out in combination with the image feature extraction algorithm to locate and track the position of the driver’s head in the video in real time,The facial feature information is transformed into array information,and the coordinates of 68 key points of face contour are detected.(2)Locate the driver’s facial eye and mouth coordinate positions according to the 68 key points of the face,calculate the horizontal and vertical Euclidean distance of the facial eye and mouth contour,set the blink and yawn threshold,cycle the video frame to count the driver’s blink and yawn times,set the fatigue threshold of the eyes and mouth through the blink or yawn times in the unit time,and judge whether the driver belongs to the fatigue state of blink and yawn.(3)Establish a spatial coordinate system,match the 2D key point coordinates of the face with the 3D key point of the face model,solve the transformation relationship between the world coordinates and pixel coordinates of the face,and calculate the transformation matrix between the coordinate transformations,namely the rotation matrix and translation matrix.Calculate the Euler angle of the head space motion through the rotation matrix,set the threshold of nodding according to the angle of head motion,and count the number of video frames of nodding,Judge whether the driver is in fatigue state.In this paper,the fatigue state of the driver is judged by analyzing the fatigue factors such as facial features and head posture.Finally,the accuracy of the test in a complex environment is more than 95%,which verifies the stability and anti-interference ability of the system.Different faces are collected for fatigue test,which can still detect the fatigue state and verify the effectiveness of the system.In conclusion,the fatigue detection system can accurately detect the fatigue state of the driver,And the algorithm has strong robustness.