Study On Driver's Eye Detection And Locating

Studying how to judge and prevent driver's fatigue is significant to reduce traffic accident. The visible expression of driver's fatigue is the fatigue state of eyes. So, how to judge the fatigue state of driver's eyes rapidly is important. Therefore, how to improve the accuracy and speed of locating eyes is mainly studied.The mechanism, possible reasons of drive fatigue, and most people's fatigue expression are analyzed firstly. The arithmetic of driver's face detection and eye locating is studied and improved. A new eye locating method based on skin color detection is put forward. It uses morphologic operation to optimize possible eye areas gradually in RGB chrominance space. Then, reduce the candidate area of the eyes by restricting every area's size and the ratio of eye's height and width. Thus, driver's face detection and eye locating is solved.This new method has the following improvement and innovation. Firstly, by usingRGB chrominance space to do skin color segmentation can extract facial features more effectively and rapidly. After the skin color segmentation, remove the noise of the object area to reduce possible eye areas directly. By masking no-eye areas and removing border-connected area, eyes' candidate areas can be decreased greatly, and the detection becomes speedier. New restrictions are added to increase the locating accuracy. The test result proves this method can locate eyes simply and rapidly. What's more, it can decrease the effect of complex background in the cab, face gesture, expression and skin color. It cannot be affected by background, skin color or gesture. For faces with some rotation angle, eyes can also be located exactly.The emulation achieves every function of the arithmetic using MATLAB language. The result of the experimentation accomplishes the destination and proves the availability of the method.