Based Perclos Driver Fatigue Detection System Design

This paper introduces a fatigue detection system based on the value of PERCLOS,expatiate the physical reason of the driver fatigue and the detection methods researching by the world now.Based on these,introduces the theory and methods of eye detection,and designs an eye detection system which based on infrared light source and red-eye effects.The system includes facial image collection,transmission and processing, to judge the state of the eyes on the basis of achieving high quality facial image, and we can get the value of PERCLOS bye statistics to make fatigue judgment.In the system,the part of facial image collection are composes of a fit of infrared LED light sources and an OV9120 CMOS image sensor.We can get better facial image by designing a special structure of the light source.USB2.0 is applied in the transmission of facial image,and the CY7C68013 chip of FX2 series manufactured by Cypress Semiconductor is chosen to implement the transmission.Threshold segmentation is used in image process and PCA is used to make recognition of the eyes state,then Kalman filter is used to track the eyes effectively and timely.This paper is divided into four sections to detail the contents of the issue. The first part introduces the physical reason of drive fatigue and the methods of fatigue detection.The second part discusses the eye detection theory and introduces the structure of the human eye detection system.The third part is the illustration of the facial image acquisition system based on CMOS and USB2.0. Recognition of the eyes state and judgment base on PERCLOS are detailed in the forth part.Summary and expectation are proposed in the end.