Research On Fatigue Driving Detection Method Based On Video

With the improvement of people’s living standards,cars have become a means of transportation for people to travel.While enjoying the convenience brought by cars,people are also troubled by the safety hazards brought by cars.It is of great significance to accurately determine the driver’s fatigue state and give an alarm during driving.In this paper,the main contents are divided into low illumination enhancement processing,driver face positioning processing and driver facial feature point extraction.The driver’s head posture angle is obtained by solving the specified feature points.Based on the eye,mouth corresponding feature points and head posture angle,judge whether the driver is in fatigue state.The specific work and innovation of this paper are as follows:The video image is preprocessed.The Retinex illumination algorithm is introduced for low illumination,and an adaptive correction function based on gamma function is constructed.This algorithm can ideally solve the problem of image brightness in the low illumination region,and take into account the color information of the image,which has certain practicability.In the face localization algorithm,this paper improves the face localization algorithm based on the combination of histogram of oriented gradient and support vector machine,and proposes the HOG feature extraction face localization algorithm based on information entropy weighting.Simulation results based on Celeb A face database show that the proposed face location algorithm has higher accuracy than traditional detection algorithms.For the extraction of facial feature points,the ERT algorithm is used for modeling analysis on the ibug-300 w dataset.By adjusting the modeling parameters,the facial feature points are more accurate.After getting the facial feature points,in order to get the head posture prediction,Zhang Zhengyou calibration method is used to calculate the camera internal parameters and distortion coefficient.Finally,the data of facial feature points and camera parameters are substituted into EPNP algorithm to obtain the conversion relationship between three-dimensional space and two-dimensional space,and then the driver ’s head posture information is obtained.In order to accurately and comprehensively detect the driver ’s fatigue driving state,different algorithms are used to solve the data of eye feature point,mouth feature point and head posture angle,which reduces the missed detection rate to a certain extent.Using Python programming language,based on PyQt5 and other development tools,a PC-based fatigue driving detection software is designed to verify the algorithm in this paper.