Yawning discrimination method based on active infrared video

Abstract

The invention discloses a yawning discrimination method based on an active infrared video. The method comprises the following steps that step one, a video frame is acquired through an active infrared camera and a face image is detected; step two, the mouth movement area is defined in a face detection box; step three, Prewitt and Canny fusion edge detection is performed in the defined mouth movement area, and the fusion edge longitudinal projection ratio R<y> is calculated; step four, the threshold T<R> of quasi yawning is determined; and step five, yawning discrimination is performed. The method is characterized in that the mouth movement area is defined in the face box, and then statistics of the Prewitt and Canny fusion edge longitudinal projection ratio in the defined area is performed to judge the yawning state so that the discrimination reliability is high.

Description

technical field [0001] The invention belongs to the field of computer vision, in particular to a yawning discrimination method based on active infrared video. Background technique [0002] In recent years, the proportion of fatigue driving in traffic accidents has been increasing year by year. Therefore, the research on driver fatigue state monitoring is of great significance in preventing traffic accidents. In driving fatigue state monitoring technology, the accuracy of monitoring technology is particularly critical to the reliability of fatigue monitoring. Although there are a large number of monitoring technologies based on objective judgment and subjective measurement, most of them have problems such as low accuracy, poor practicability, complicated operation, high requirements for monitoring equipment, and weak anti-interference ability. Research shows that the fatigue state discrimination method based on subjective measurement of facial features has the advantages of ...

AI Technical Summary

Problems solved by technology

In order to adapt to all-weather and continuous monitoring and reduce the impact of light, it is proposed to use an active infrared camera to monitor the driver's yawning state, but in the infrared video, there are two problems in yawning discrimination: one is the positioning of the mouth area

Generally, the degree of opening and closing of the inner contour of the mouth (height-to-width ratio of the lip area), the area of ​​the lip contour, and the lip dynamic matching model based on gray energy corner points are used as the basis for judging yawning. Since regional planning is used in the mouth positioning link method, then the method based on aspect ratio and area statistics is not applicable, and even if the aspect ratio, area statistics and dynamic matching model are used after the mouth is detected in the visible light video, since the size and angle of the face in the video change at any time , its reliability is not high

Images