Design And Implementation Of Fatigue Driving Identification Based On Facial Features

Fatigue driving is one of the main causes of road traffic accidents and is a major problem affecting people’s safe travel.In order to reduce traffic accidents caused by driver fatigue,so the research on fatigue driving detection will be carried out in depth.In this paper,after introducing the existing fatigue driving detection methods,a fatigue recognition method based on facial feature fusion is proposed,and the main research works are.1)Image pre-processing.As the images are subject to light imbalance and various noise disturbances in the process of acquisition or transmission,the images are subject to smoothing and noise removal and light equalisation processing,so as to enhance the effective image information and ensure the accuracy of face detection.2)Face detection and key point localisation.The AdaBoost algorithm based on Haar-like features is used for face detection,and combined with the integral image method to improve the calculation speed and exclude most of the background effects,so that the key point detection range is greatly reduced.The cascade regression method is used to locate the 68 key points of the face,and the eyes and mouth can be quickly located according to the position of the key points,so this method has strong anti-interference performance.3)Recognition of the state of the eyes and mouth.Among the 68 key points of the face obtained,the information can be located to the human eye and mouth,and then the aspect ratio of the human eye is used as the input of the SVM classifier to recognize the open and closed state of the eyes.The mouth is then located according to the position of the key points,and the mouth aspect ratio MAR is calculated to determine the state of the mouth.4)Fatigue determination.Four features were extracted when the face was fatigued:blink frequency BF,the longest duration of continuous eye closure MECT,the ratio of the number of frames closed based on PERCLOS ECR,and the number of yawns.Based on these four fatigue features,a fatigue recognition method based on a combination of several features is proposed,and the driver’s fatigue status is judged comprehensively by these indicators and prompted accordingly.This thesis uses PyCharm provided by Jetbrain as the platform with OpenCV computer vision library for Windows development,using Python to write the program.The system is works in real time and provides an accurate picture the real situation of the driver and provide appropriate alerts if fatigue is found to exist.