FPGA Design And Hardware Implementation Of Driving Fatigue Detection System

The traffic accident statistical analysis table of China shows that900,000people inour country died in traffic accidents from2001to2011. According to the NHTSA data,about33,561people were killed in traffic accidents in2012. The traffic accidents can betraced to various causes, including fatigue, speeding or overloading, of which, thefatigue driving is the primary cause of traffic accidents. Scientific research suggests thatreducing fatigue driving can sharply reduce traffic accidents, and it’s significant to curbtraffic accidents.In this thesis, taking current methods both at home and abroad into account, wedesign a fatigue driving detection system with high performance, practicability andstability. This system uses ADV7183B and XC6SLX75as core controller, which canmonitor fatigue in actual driving, reduce traffic accidents and give drivers andpassengers better protection.In view of the "bright pupil" effect that caused by the human eyes in the850nminfrared light irradiation, we propose a method which using image processing techniquefor driver fatigue detection system. This system consists of three modules: the front-endimage acquisition module, on-chip system module in the middle as well as the back-enddisplay and alarm module. In order to get a better infrared image, the opticalinterference filter we used in the system is91.97%peak transmittance and the centerwavelength is850nm, with proper infrared source to increase the effect of CCD sensor.We utilize improved simplified PCNN (Pulse Coupled Neural Network) model in thissystem and real-time processing in the chip of FPGA conducted by VHDL to achieveaccurate segmentation of human eye "bright pupil". Meanwhile, the Perclos algorithm isalso used to detect the state of fatigue of drivers. Finally, the state of human eyes andthe final results are displayed on external display module in real time, giving a warningor other emergency methods to drivers in fatigue driving.We design a non-contacted fatigue detection system by using image processingmethod based on behavioristics on the platform of FPGA. Lots of experiments showthat this design can realize real-time, accurate and stable fatigue detection. Furthermore,this system has high integration and can be transplanted easily, providing referencevalue for other related designs.