Design And Implementation Of Drivers' Fatigue Detection System Based On FPGA

Motor vehicle which as the product of the progress of human civilization and society has made great contributions to the economic development .However, with the increase of motor vehicles and transportation development, traffic accidents occurs more frequently,which bring people's production and life serious impact.Both domestic and foreign, in the traffic accident statistics due to driving fatigue losses of life and property of the Numbers are astounding.Therefore, how to effectively monitor and prevent driving fatigue, to reduce traffic accidents has the extremely vital significance.The main objective of this subject is to develop a set of FPGA-based natural light, driver fatigue detection system.Comparing the current domestic and international popular detection methods' key technology and research result, the proposed driver fatigue detecting method is based on computer vision, non-contact, non-physiological load. Using FPGA logic resource design image data acquisition module, vga display module and image processing module .Firstly ,use face-color model to detect and locate face region , secondly,aiming at the eyes location,the paper bring the mothed of geometric features ,integral projection ,region growth to locate the eyes.finally through the eyes of the state of fatigue with PERCLOS metrics and frequency of wink to determine whether the driver fatigue.In this paper the HW-SW codesign flow is used,the system is based on Xilinx's Virtex-II Pro FPGA, in the chip there are two IBM PowerPC405 processor.assign proper work to hardware and software,make good combination of FPGA's flexible logic resource design and powerful processor software functions, assign part of the video image processing to hardware greatly reducing the burden of embedded processors,assign other procedure that not easy to process by hardware module to PowerPC by write C code.Experiments result shows the hardware and software efficient cooperation. make system achieve high speed and performance, this system has high real-time performance and accuracy.