A Study Of Driver Fatigue Detection Based On Computer Vision

Driver fatigue is one of the chief causes of traffic accident, and driver fatigue detection is attractive in Intelligent Transportation Systems (ITS). This paper presents a survey of driver fatigue detection, and chooses computer vision based method as the study target to build a real-time detection system of driver fatigue. The system has a CCD camera which collecting videos of driver's upper-body to a computer, and uses some algorithm of computer vision to compute the percentage of eyelid closure , and then, estimates the level of driver fatigue . The core content of this paper is that realizing and improving algorithm such as face location, face tracking, eye location, eye information extraction and fatigue level calculation.A method based on skin color segmentation and face verification is used to locate face in an image. In order to do this, we convert raw video images from RGB color space toYC ; and then, segment out each skin color region from the background; finally, bind the face region by a rectangle box according to some prior information. A method, named as Shrink Shear (S-S), is proposed to correct the results of face location for the sake of improving anti-jamming ability of the face location algorithm. bCrA color-based tracking algorithm, CAM Shift, is chose to tracking face via a flesh color model, and some improvements on the tracking process are made. Face location algorithm was executed before tracking face. A piece of color image in the face region located before was use to build the color model of the tracking target, and the location of face region was used to initialize the searching window of CAM Shift. Finally,S-S algorithm is used to correct the result of CAM Shift.In order to locate eyes on the face, a Cascade Classifier based on Haar-Like features is used to search all existing eyes in the face region in different scale. Eye tacking is aiming at eliminating the noises and interferences of the environment. In order to do this, a series of strategies are designed specially to track lost eyes. Firstly, the eye losing condition can be classified into two categories, single eye lost and two eyes lost. And then, each lost eye can be located according to location of face and historical information of eye location.Eye information extraction is mainly means to compute percentage of eyelid closure. Some relative researches have been analyzed, before a novel method, named as iris's external rectangle based method, is proposed to calculate the percentage of eyelid closure. The principle of the iris's external rectangle based algorithm is that: bind the iris region by using an external rectangle, use the ratio of the width and height of the rectangle to estimate the percentage of eyelid closure. This method can represents eye states more precisely than others relative method.Finally, PERCLOS algorithm is used to evaluate the fatigue level of driver. The results of tests based on PC and DSP system show good ability of this research.