In this paper, in order to implement a computer vision-based recognition system of driving fatigue. In addition to detecting human face in different light sources and the background conditions, and tracking eyes state combined with fuzzy logic to determine whether the driver of the physiological phenomenon of fatigue from face of detection. Driving fatigue recognition has been valued highly in recent years by many scholars and used extensively in various fields, for example, driver activity tracking, driver visual attention monitoring, and in-car camera systems. In this paper, we use the Linux operating system as the development environment, and utilize PC as the hardware platform. First, the system uses a camera to obtain the frame with a human face to detect, and then uses the frame to set the appropriate skin color scope to find face. Next, we find and mark out the eyes and the lips from the selected face area. Finally, we combine the image processing of eyes features with fuzzy logic to determine the driver's fatigue level, and make the graphical man-machine interface with MiniGUI for users to operate. The results of experiment show that we achieve this system on PC platform successfully.
