Research On Driver Fatigue Detection Algorithm Based On Euler Image Magnification

The popularization of the car has brought many conveniences to people’s life, but many security problems occurred during the process of car’s moving. According to the worldwide traffic accident data analysis, driving fatigue has become one of the main causes of the traffic accidents. In order to reduce the incidence of traffic accidents caused by driving fatigue, many experts and scholars respectively do lots of research about the drivers’ physiological characteristics and vehicle characteristics of driving fatigue detection. This paper applies the amplification of video data by Eulerian Video Magnification to the driver fatigue detection.Eulerian Video Magnification is studied in the driving fatigue detection. Firstly, this paper introduces the concept of fatigue driving as well as analyses its formation process and influence factors as well as researches on commonly used non-contact detection method systematically. Secondly, this paper researches on decomposition and reconstruction of the image of pyramid and some based on wavelet transform. The principle of Eulerian Video Magnification is systematically studied in this paper. The algorithm is realized by the color space conversion, spatial filtering, temporal filtering, amplification and synthetic image of video frame sequence diagram. The algorithm can amplify the motion and color and give an analysis of sample.Finally, the enhancement of Eulerian Video Magnification for the fatigue detection is studied. First I collect multiple sets of experimental video and use the algorithm to deal with the video and choose the best magnification and filtering range of fatigue characteristics(eye features, head feature and heart rate signals e.g.). Then, I process the multiple sets of video using the best magnification and filtering range. Last this paper verifies Eulerian Video Magnification can enhance the effect of fatigue detect by face detection and fatigue characteristic extraction.