Fatigue Driving Detection System Design And Implementation Based On Human Eye Feature

With the improvement of people’s living standards, every year a large number of people want to buy car, accompanied by traffic accidents has also increased. Therefore, if we can develop a rapid and accurate fatigue detection system that will protect people’s lives and property safety significance.This article is a summary of previous research work, based on the detection of fatigue in the face detection, eye location, eye fatigue state identification and determination methods such as in-depth research and design implementation. Design characteristics of the human eye by the driver fatigue detection system based on video capture module, image pre-processing module, face detection module, the human eye positioning module, the human eye state recognition module determines fatigue and fatigue alarm module.(1) Video capture and image pre-processing module: In this paper, the video image to obtain driver information through the camera. Image processing technology, the acquired video image graying, seeking histogram, histogram equalization and median filtering pretreatment.(2) Face detection module: In this paper, to realize the detection of the driver’s face based on AdaBoost cascade algorithm Haar feature, the face detection algorithm for high speed, good effect and so on.(3) Human eye positioning module: This paper uses two methods targeting the human eye, the first gray projection method to the human eye rough location, and then use the complexity of the analysis block method points to the human eye for accurate positioning.(4) Human eye state identification module: In this paper, the mean projection and variance projection method to achieve the status of the human eye(eyes open, eyes closed) judgment, to lay the foundation for the subsequent determination of fatigue.(5) Determination of fatigue and fatigue alarm module: In this paper, we use PERCLOS guidelines to achieve eye fatigue determination. Speech recognition system using Microsoft Speech SDK voice alarm function.In this paper, I use Visual Studio 2010 as development environment and use VC++ as development language, real-time using OpenCV. On the PC, robustness and accuracy of detection systems to do the relevant tests, the system can achieve fatigue detection function, and in the uniform illumination, correct posture, do not wear glasses under conditions with high accuracy.