Research Of The Fatigue Warning System Based On The Eye Movement Detecting

In today’s high-paced life and efficient work, fatigue is one of the important factors that lead to make accidents, such as transportations, health care, etc. It is becoming a hot research that how to prevent and reduce the occurrence of the similar accidents. It can effectively prevent the happening of the accidents for the ringing when the drivers just appear fatigue with monitoring the driver’s fatigue.It is the study of motion that to detect a kind of human eyes fatigue warning system. It is a system to collect real-time video image data processing, to extract of the human eye fatigue data, and then to judge the fatigue of the driver with early warning work based on ARM+ Linux embedded system platform. It is the hardware platform which is designed with a series of module involved of ARM development board based on S3C2440, USB camera, LCD. It is constructed fot the running environment of system software with the Linux system and OpenCV computer vision library on this platform transplantation. It is the determination of the functions program designed which are completed such as the video image acquisition, eye recognition of the human face, the eye fatigue feature extractionFirst of all it was related fatigue algorithm based on human research and simulation in PC. It is adopt the method of cascade classifier Haar features, to detect the human face and the eye area in video image with a USB camera collection, to track the eye area based on the frame difference method and the location information of human eyes in face method, to extract the human eye fatigue feature, to determine the driver’s fatigue state combining of blinking eye fatigue, close your eyes time and Perclos decision mechanism.And then designed a ARM+Linux embedded system platform, and transplant the embedded Linux operating system, Qt graphical interface library and OpenCV library to build a computer vision development environment for system software on this platform, with the development of the system application. And it was tested and validated through the experiment on the design of platform and the algorithm.