Study On Ship Pilot Fatigue Fetection System Based On Face Recognition

Work at sea is characterized by high intensity,long hours,and disrupted rhythm of day and night,which will cause the ship’s pilot to feel fatigue due to the large physical and mental load.In the state of fatigue,the driver will have a lack of concentration,slow movement and other behaviors,which can easily lead to maritime traffic accidents.Therefore,it is necessary to develop a system to monitor the driver’s fatigue status and alarm in real time,which can effectively stop the driving behavior of drivers,which is important to reduce maritime traffic accidents and guarantee the safety of people and ship navigation.At present,the fatigue monitoring technology applied on land has been developed maturely,but the fatigue detection system used in maritime environment is still in the exploration stage,which cannot avoid the influence of factors such as maritime light,weather and ship sway on the accuracy of image recognition.To solve the problem,we propose a ship driver fatigue detection system based on face recognition.The main research contents are as follows:(1)Combining the advantages and disadvantages of existing fatigue detection techniques on land,we study a face recognition model based on deep learning YOLOv5 s lightweight network structure,which can effectively solve the problem of low face recognition rate in complex maritime environments and improve the accuracy and speed of face detection.(2)We study a classification model based on CNN convolutional neural network to obtain facial feature information and locate the key points of five senses organs by face recognition,and intercept the key areas of five senses organs of the face by using three chambers and five eyes in two equal parts to achieve accurate classification of the open and closed state of the eyes and mouth.(3)In this paper,we use the PERCLOS fatigue judgment criterion to improve the accuracy of fatigue detection by calculating the blink frequency and yawn frequency based on obtaining the eye-mouth opening and closing state,combined with the new index of eye gaze frame.In this paper,a complete system of ship pilot fatigue detection system is designed and the related experiments are studied and discussed.Tested on the Yaw DD yawning video dataset and the actual scene dataset simulated in the laboratory,the fatigue recognition accuracy reaches 96% and the speed reaches 76 frames per second,and the system achieves optimal accuracy and real-time performance.The experimental results show that the fatigue detection method can effectively detect the fatigue state of ship drivers and alert them in time,which is creative and practical.