Research Of Fatigue Driving Detection System Based On Infrared Video

Long time continuous work result in the driver’s visual nervous, and prone to a fatigue state. In the process of driving, if the driver fatigue, his nerve is easy to lose vigilance,however, the driver’s briefly failures could cause a major disaster. Therefore, study fatigue detection system is very important. According to the current website statistics released by the national, in each year our country involving about 200000 casualties in the traffic accident, and the fatigue is the main factor of causing traffic accidents. Among them, in the traffic accident statistics shows that 31% of freight truck driver death in traffic accident directly, while 58% of freight truck safety accident is caused by driver fatigue driving.Therefore, study of fatigue testing device has an important significance and practical application value. In recent years, fatigue driving detection has become a hot research problem in machine vision applications, especially for the night image research is increasing.This thesis first introduced the background of this topic research, expound the significance and the present situation of domestic and foreign scholars study of fatigue detection. Then based on the demand of driver fatigue detection, we design a driver fatigue detection system based on TMS320DM6437 video development platform. On the first, we proceed image denoising processing. On the second we gathering the pilot video image in the system adopted AdaBoost algorithm of face detection method based on skin model have on face detection, and according to the integral projection method combined with AdaBoost algorithm of face region to conduct eye location; Then with the improved MeanShift algorithm to achieve real-time tracking of the eyes, and to extract eye feature; At the same time, combining with the state of fatigue to judge different drivers’ eyes state:completely open, closed completely or half open and half closed. Finally, the alarm system will detect driver fatigue state, and timely remind the driver.The fatigue test system consists of image acquisition, image processing, fatigue judging module and alarm system. Using infrared illumination of CCD cameras collect driver’s facial image video, and convert this facial image video into digital image signal by TVP5150 video decoding chip, after that we get on image processing and fatigue judging work, according to the change of state of the eye, select it and use it as the basis of whether the driver fatigue, when the driver appeared fatigue status, the system will give alarm in time.The system selects the active infrared night vision technology of infrared CCD camera as a video capture device, which is does not interfere with drivers normal driving,the system can not only realize continue real-time monitoring to a target, but also not interfere with the system operation under the condition of dark or the car has in lighting situation.