| In today's growing motor vehicle, the vehicle becomes our indispensable transport tool, followed by traffic accident is unavoidable. And fatigue has become "number one road killer" in the traffic accident. Therefore, it is very meaningful to judge whether the driver fatigue driving and alarm.Paper extracts the characteristic parameters of eyes and mouth to detect the fatigue state. The main work of the paper includes the following. First, face detection. In paper, we use the HSV color space which is not affected by the facial expression, pose and angle to quickly locate the human face, and get the initial position of the human face image. At the same time, we use the skin color model to detect the face region to provide the initial position for ASM to solve the difficult problem of ASM initial alignment. Secondly, the ASM algorithm is used to track the eye and mouth area of the eye and mouth. In order to improve the accuracy of the key parts of face, paper improved AAM algorithm, using 48 points model, and combined HSV and local AAM algorithm to reduce the search time and improve the robustness of face localization. Again, the eye state judgement. In order to more accurately determine the eye state, we choose the Canny operator to extract the contour of the eye. Then, the distance between the horizontal line segment of the upper eyelid to the inner and outer corner of the eye is chosen as the basis for judging the degree of eye closure, calculating the PERCLOS value, and judging whether the eye is tired. And after all, the mouth state judgment. Local AAM using 19 points of the mouth model, to provide the region more accurately enough to carry out the mouth and mouth open degree judgment. Then calculate the PMECLOS.Finally, paper presents a fusion of eye fatigue parameters and fatigue parameters are given and the mouth, the driver's mental evaluation model. The single fatigue parameters into two fatigue parameters, improve the fatigue detection accuracy, and get better results. |
PDF Full Text Request