Research Of Driver Fatigue Monitoring Method Base On Facial Features

A recent study showed that the cause of the accident25to30%is caused byfatigue driving. The traffic accident caused by fatigue driving would lead to hugeproperty losses and casualties every year. Many countries are actively involvedin research work related to the field of fatigue driving detection. the purpose of thestudy is to reduce the traffic accidents caused by fatigue driving as the starting point,so the research has important social significance.This design is to use the information change of driver’s facial features in thecourse of driving to complete the detection system of fatigue driving. the feature offace images by captured is extracted, analysis and comparison of the facial featurechanges. Through the pixel area changes of human’s eyes and mouth size to judgewhether the driver is in a state of fatigue. The AdaBoost algorithm base on haarfeature is used to analyze frame images to locate the human face, the frame images arecaptured by camera in real time. At the same time of location, based on thedisplacement characteristics of driver setting where the failure range of tolerance tocontrol jitter effects, retain data effectively and eliminate invalid data. The feature ofhuman’s eyes and mouth region is extracted by haar rectangle. For initial location of thehuman’s eye and mouth region using improved Adaboost iterative algorithm to locateprecisely. The area of human eyes and mouth is saved and calculated in real-time byusing the moving average smoothing algorithm, which is based on stack.According to the pixel area change the human’s eyes and mouth caused byblinking, yawning fatigue behavior to judge whether the driver is in a state offatigue. When the driver closed his eyes over the threshold or yawn,considered in a state of fatigue and feature display prompts warning.The design of the system using Microsoft Visual Studio2010developmentplatform, through the VC++programming to realize the algorithm. In the design Implementation uses the Intel OpenCV library. The experimental results showthat, the algorithm can quickly and accurately detect the fatigue state of driver whenthe blink of an eye, yawning.