Study On The Detecting Method For Driver's Head

Safety, traffic jam and environmental pollution are difficult problems of international vehicle Transportation field. Especially safety problem in transportation field is paid more and more attention worldwide. Driver's human factors have been one of the most important causes of road accidents. Safety Driving Assist technologies can support greatly for reducing the road accidents due to drivers'human factors. Driver monitoring has been a focus of Safety Driving Assist technologies research. Driver's fatigue and distraction warning system take important role on reducing accident rate.Drivers fatigue judge by detection based on driver's face organ. Many research have focused on monitoring the driver's face, eye, pupil and so on to obtain his/her face rotation and orientation, eye activities, eye blinking rate, gaze direction, finally to determine his/her fatigue or distraction state. To detection of the face organ, driver's head must locate accurately. It will be provide a good foundation for face organ detection, and monitor drivers in the course of driving behavior better.Study on driver's fatigue detecting method and exploit driver's fatigue and distraction warning system is adapt development of vehicles technical. A method which uses machine vision to monitor driver's fatigued driving behavior in real time. From the information we can estimate safety of vehicles. It can reduce road traffic accidents caused by fatigue driving. This system has important engineering and science significance. Therefore, the research of this aspect has been a studying focus of international experts and scholars.The research work in this paper can be categorized into four parts, i.e. driver's face skin division, based on ellipse fitting driver's face detection, driver's hair contour detection, and based on curve fitting driver's hair detection.In the aspect of driver's face skin division: Firstly, input color image use lightcompensation to eliminate color bias. Secondly, input color image is segmented skin-like and non skin-like regions in YCrCb and HSV space separately, then using logic"and"operation to combine them. This method can improve the efficiency of skin color segmentation. Experiment results show this method can segment skin-like and non skin-like regions effectively.After face skin division based on YCrCb and HSV skin color model the combined image appear noise. Using mathematical morphology and median filtering methods take out noise. Then detect face contour using face contour tracking algorithm. Finally fit an ellipse based on face contour detection. Ellipse fitting method is different from the quadratic curve fitting method. The quadratic curve fitting method relies on good data distribution or a lot of calculations. The methods in this paper provide the best compromise on fitting speed and accuracy. And the method theoretically guarantees only the ellipse fitting. Experiment results show that this method can effectively fitted face contour.In the aspect of driver's hair detection: firstly, input color image transform V space image of HSV color space. Secondly, the binary image is achieved based on Otsu algorithm. after detecting hair region, By analyzing Sobel operator, Laplacian operator, Canny operator edge detecting results, Canny operator can be found more edge points, thus driver's hair contour detect based on the Canny edge. Then separately find the outside and inside hair edge points.After find the outside and inside edge points, the paper use the least square method and third uniform B-spline fitting driver's hair contour. The least square method fitting hair contour is smoother. Third uniform B-spline method fitting hair contour more accurate than the least square method. Experiment results show that this method can accurately fitting hair contour.The software application system is developed using Visual C ++6.0,Matlab7.01and OpenCV. The arithmetic methodologies have been tested and reached the prospective targets, and the availability has been probed.