Design And Implementation Of The Human Eye’s State Detection Based On Omap4460

The occurrence of traffic accidents are many caused by the driver’s fatigue, so the realtime detecting the driver’s state of fatigue can largely reduce the number of traffic accidents, and determining the fatigue status can be judged by the human eye’s state. The state of the human eye can be the opening of the human eye and eye open frequency. Detecting eye open and close state is the foundation of human eye frequency detection, so it is of vital importance in the whole detection system. And the human frequency of eye open in sober and stuck being significantly different conditions can be an important basis of detecting the state of fatigue. There are many scholars on the driver’s fatigue state detection system based on the human eye state detection to do a lot of research, but these systems are which there is a lack of testing precision and real-time performance, as a result of the detecting speed limit many systems are only based on PC, this makes the detection system is difficult to implement in practical application. So this article puts forward the human eye state detection based on Omap4460, which is based on arm, the integrated system of small size, low power consumption.It provides a possible implementation for the human eye state detection of application.Combining with research status at home and abroad, this paper set the human eye detection test platform based on Facedetection module of Omap4460 for omap Linux kernel which realized the human eye state detection based on Adaboost and Cam Shift. Because the resolution of the actual simulation PAL video signals being 720x576, the scheme which this article put forward can achieve a resolution of 720x540, frame rate of 25 fps eye’s state detection. Specific work content is as follows:1.In SOC because no Omap4460 Facedetection module in the Linux kernel driver, therefore this article developed the driver of Omap4460 based Facedetection module.According to it and in order to be combined with Open CV, this paper also developed a Facedetection interface based on Open CV and Omap4460.2. This paper adopted Adv7280-m conversion chip, thus it developed a device driver on the chip and developed Gstreamer v4l2 src plug-in based on the driver, by which the video data package work was achieved.3. It proposed a detection algorithm based on the eye contour area of eye open and close state detection algorithm and implemented eye open frequency detection based on open and close state detection of the human.4. In view of the human eye state detection, this paper developed a fdifscale plug-in, in which the algorithm of the human eye state detection and tracking was encapsulated. Combining with the other video processing module in Gstreamer, the plug-in realized the real-time detection of the human eye condition.5. This article put forward the real-time and accuracy of the human eye state detection scheme to test the experiment and summarized the result.