Research On Human Fatigue Recognition Of Highway Tunnel Drivers

With the rapid development of highway tunnel construction in China,Traffic accidents are becoming more and more serious.One reason of road tunnel accidents is fatigue driving.Useing fatigue monitoring methods to identify fatigue driving has been more and more important.In the tunnel,the driver’s eye fatigue problem is more serious than on the highway.Thus,the possibility of a traffic accident in the tunnel is higher.In order to classify driver’s eye fatigue,this paper conducted a study on driver’s eye fatigue recognition based on PERCLOS method.The main contents are as follows:(1)This paper analyzed the existing face data set,designed a tunnel driving experiment,and built A face data set for highway tunnel drivers.(2)This paper analyzed the existing face and eye recognition algorithms.For tunnel driving video,this paper recognized the driver’s face and eyes based on the Haar feature in the Adaboost algorithm.In addition,this paper built a driver’s eye fatigue data set based on driver’s eye recognition results.(3)This paper designed a integrated convolutional neural network model,and realized eye’s classification and recognition.In order to improve the accuracy and efficiency of eye’s classification,this paper used the ANN and CNN networks to build a ICNN network model,which maked the result of human eye classification better.(4)This paper calculated the PERCLOS value of the driver’s eye,and improved the PERCLOS algorithm by combining the blinking frequency and the maximum eye closing time.Based on the improved PERCLOS value,the fatigue rating is divided.In this paper,the driver’s fatigue rating was divided based on the improved PERCLOS value.This paper used the ICNN network model and PERCLOS results to achieve the recognition of driver’s eye fatigue in Highway Tunnel,which has important significance in helping traffic management departments manage traffic,reducing the incidence of traffic accidents and improving traffic operation safety.