Design Of Fatigue Driving Detection Algorithm And Reminder System Based On Deep Learning

With the increasing degree of social industrialization and the establishment of a comprehensive well-off society,people's lives are more convenient and travel options are more diverse,and because of its convenience,cars have become the first choice for more people to travel.It brings great convenience to people's travel,but it also brings potential safety hazards while bringing convenience.The frequent occurrence of automobile road accidents will have a huge impact on social generation and people's lives.Fatigue driving occupies an important position in many factors that cause traffic accidents.Carry out testing to solve the harm caused by fatigue driving fundamentally.At present,there is no fast,convenient and accurate fatigue state detection method to detect the driving state of the driver.Therefore,designing a fast,convenient and accurate fatigue detection algorithm is of great significance to social production and people's lives.In order to reduce the social harm caused by fatigue driving,this article aims to design an accurate and convenient driver fatigue detection algorithm,and design a fast,convenient and accurate fatigue detection system based on the algorithm.The system uses the currently popular deep learning method to detect the state of eyes and mouth in the face through face detection and face key point network,and then determine the fatigue state of the driver by comprehensively judging the state of both eyes and mouth,and Based on the algorithm,a system is designed to verify the feasibility and accuracy of the algorithm.The main research work of this paper mainly includes the following aspects:(1)Research and analysis of the development of fatigue detection related topics,and understand that the commonly used fatigue driving detection index is based on the traditional image processing method PERCLOS algorithm,which uses computer vision to calculate the proportion of eye closure time to determine whether the driver is In a state of fatigue.(2)A brief introduction to the basic principles and basic structure of neural networks in popular deep learning,a general understanding of neural networks,and several classic neural network structures and their respective applicable scenarios.(3)Introduced the representative network structure of fully convolutional neural network and fully convolutional network YOLO-v3,and using YOLO-v3 as a face detector,designed a person for fast and accurate detection of the driver's face Face detector,and use this detector in the driver fatigue detection system designed in this paper.(4)Introduce some neural network structures that pursue fast and light weight,analyze the principles of deep separable convolutions and their application in the Mobile Net-v2 network structure,and design a network for fast based on the Mobile Net-v2 network structure A face key point detector that detects the position of key points in the face,which is an important part of the driver's fatigue driving detection system.(5)A complete set of driver fatigue detection system is designed,which combines face detection,face key point detection,driver fatigue state judgment and other modules,and verifies the usability and reliability of the system through experiments.The results achieved in this paper are based on the YOLO-v3 face detection network and Mobile Net-v2 face key point detection network two deep learning networks to build a fatigue driving detection algorithm,and on this basis,a simple fatigue driving reminder is designed system.The contribution of this design to the society is to explore the effect of deep learning methods on fatigue driving detection,and to pave the way for more mature technology applications in the field of fatigue driving detection.The future development in this article means combining theory with practice and trying to apply new knowledge and new technologies to areas closely related to people's production and life.Technology serves people's lives and promotes social progress.