The invention discloses a
rail traffic driver fatigue state monitoring method. Through collecting the
head neck image of a driver, the position of a face is determined, the
facial expression image of the driver is collected according to the face position, and the positions of eyes and a mouth are determined; Left and right
eye state characteristics and a mouth expression characteristic are recorded and are transmitted to a state identification unit, the characteristics are classified by a classification
algorithm, and whether the driver is in a drowsy state or not is judged; According to the face position, a coordinate
system is established at the face, a coordinate
system movement characteristic and the
head and neck movement characteristic in a
human body drowsy state are compared, and thus whether the river is in a drowsy state or not is judged; and if a condition that the driver is in the drowsy state is detected, the
alarm device emits an alarm prompt. The method has the advantages of a simple and efficient
system design, high detection precision, low requirements of
image processing and corresponding hardware, low cost, small area occupation, and strong robustness, the fast and correct fatigue driving judgment can be realized, the detection sensitivity is high, and the method is safe and reliable.