Research Of The Fatigue Warning System Based On The Detection Rate Of Eye Movement

The accurate detection of iris designed a driver fatigue detection and warning system. This design in the research process is mainly aimed at the existing algorithms to improve,the fisrt is to achieve more rapid and accurate identification of extracting eye driver fatigue,and combine real time monitoring of the mental state of the driver during driving through the fatigue state information of the eye to achieve a complete eye movement rate detection based on the fatigue driving condition monitoring systemand validation and analysis of the effects of system testing.The first is fatigue warning procedures related to verification by PC, then use based on ARM+Linux platform for fatigue early warning plan implementation. The fatigue state discrimination method based on a novel for driver fatigue detection based on eye movement rate of fatigue detection method, when people are in a state of fatigue, eye movement rate would drop dramatically, according to this characteristic, combined with the processor for CortexA8 ARM chips, and the Linux operating system, Qt graphic interface and OpenCV Computer Vision Library, design of fatigue warning system based on embedded platform. The cascade classifier Haar feature call OpenCV computer image processing in the library to extract the human face and the human eye region, then through edge detection and corner detection to extract the eye iris center and fatigue characteristics of eye position and other related information, Finally based on the fatigue early warning method based on human eye movement rate.Finally, the program will validate the cross compiler development board can run into the executable file in Linux system, and transplant procedures used in the OpenCV library and Qt interface software used to support software development environment, Finally verify the fatigue warning procedure under the ARM+Linux platform.The experimental results show that the fatigue warning scheme based on eye movement rate from the real-time and accuracy obtained good results.