Study On The Detection System Of Locomotive Driver Fatigue Based On Image

With the adjustment of the layout of the railway productivity, locomotive routings continuing to extend, the phenomenon of driver's overwork are very serious. Driver fatigue caused by driving the whole railway locomotive has become an important cause of traffic accidents. Therefore, developing a intelligent fatigue monitoring system applied to train drivers plays a significant role in the safe driving of the train.This paper presents a system using a CCD image sensor to capture the driver's facial image, through analysis of the state of the human eye to determine the degree of fatigue system. Mainly referring to two indicators to determine the extent of driver fatigue: PERCLOS( Percent Eyelid Closure )and AECS( Average Eye Closure Speed). The system uses the classifier cascade of AdaBoost algorithm based on Haar features in the detection of a human face, based on a template matching method to locate region of the human eye in the detected face and detect the status of the human eye. When the system detects the driver is in sleep mode, it can alarm to remind the driver, and keep relevant data uploaded to the ground control center.In this paper , the AdaBoost cascade classifier algorithm based on Haar features and the template matching algorithm is used, according to whose principles and methods described, we carry out the software programming and their implementation, and test algorithms for achieving face detection and eye location and its status recognition through several collected test sets. Tests show that the AdaBoost algorithm and template matching method with a detection accuracy rate and running speed advantages. At last, for the lacks of the algorithm exposured in the tests, we raise a further improved method.The full text is divided into six chapters chapters. The first chapter introduces and points out the background and significance of this study, and gives an overview of the driver fatigue detection technology at home and abroad. Chapter II first determines evaluation indexes for the fatigue status of detection, and proposes the AdaBoost algorithm based on Haar features to detect the human face. Finally,we give the detailed design of the fatigue detection system. The third chapter describes the Haar-based AdaBoost face detection algorithm, including a detailed description of the basic principles and the introduction of specific implementations. In the Chapter IV the principle of template matching are introduced first,based on which,we proposed the template matching method applied to this system to detect the human eye, and discuss the key issues of template matching. Finally, the input image's preprocessing process in the location and status of the human eye is brief. In the Chapter V ,using some test sets ,we carry out the tests of two kinds of face and eye detection algorithm, and give the results and the analysis of the causes of false detection. Chapter VI Conclusion and outlook, the main research work of the full text is summarized. According to the actual fatigue test's results, we analyze the algorithm deficiencies and propose some reasonable improvement.