Driver face characteristic distance correction method based on facial expression fatigue driving detection algorithm

Abstract

The invention discloses a driver face characteristic distance correction method based on a facial expression fatigue driving detection algorithm. The method comprises: when a driver shakes up and downand shakes the head or tilts the head leftwards and rightwards, the actual value of the transverse or longitudinal feature distance between the mouth and eyes is derived by changing the included angle between the transverse or longitudinal feature line segments of the mouth and the eyes and the positive direction of the camera; a right hand space rectangular coordinate system is established withthe head of the driver as the center, the x-axis direction is the left ear pointing to the right ear, the y-axis direction is the head pointing direction, and the z-axis direction is the forehead pointing to the back head spoon. The rotation angles of the head on the three coordinate axes are obtained; wherein the pitch angle yaw represents the angle of rotation of the head of the driver around the y axis, the yaw angle pitch represents the distance of rotation of the head of the driver around the x axis, and the roll angle roll represents the angle of rotation of the head of the driver aroundthe z axis, namely the Euler angle of the head of the driver in the three-dimensional space. The obtained Euler angle is used for correcting the characteristic distance, so that the driver facial expression fatigue driving detection accuracy and precision are improved.

Description

Technical field [0001] The invention relates to the technical field of facial expression feature extraction in fatigue driving detection, in particular to a method for correcting driver facial feature distance based on a facial expression fatigue driving detection algorithm. Background technique [0002] After a person enters a state of fatigue, the face will show a variety of typical drowsiness manifestations, such as the involuntary closure of the upper and lower eyelids, the obvious increase in blink frequency within a period of time, the gradual increase in the time of each blink, the slightly open mouth, Yawn and wait. As the two most active parts of the face, the movements of the eyes and mouth can best reflect the above-mentioned changes in expressions under drowsiness. With the development of computer vision technology, face detection technology has become more and more mature. At this stage, face detection based on computer vision has reached a commercial level in terms...

AI Technical Summary

Problems solved by technology

In the above scenarios, although the ratio of the height to width of the eyes and the mouth can solve the problem caused by the distance between the face and the camera, considering the actual situation, the installation position of the camera cannot be completely facing the driver’s face, and the driving Due to complex road conditions, bumps in the driving process, or the needs of the actual situation (such as checking the dashboard or rearview mirror), the driver's head may shake slightly from side to side or up and down, so the facial images captured by the camera cannot always be same angle

Considering the 2D camera used in practical applications, as attached image 3 As shown, when the driver is facing the camera, the length of the line segment is exactly the feature distance that needs to be extracted, and when the driver’s head is not facing the camera, due to a certain deflection angle of the face relative to the camera, the image captured by the camera at this time is There is a deviation in the feature distance, so the fatigue features calculated according to the fatigue feature extraction method proposed above are not accurate enough

Images