Driving Detection Methods, Based On Face Detection Technology Fatigue

With the increasing number of the vehicles, more and more traffic problems have been concerned. One of the main cause, driving fatigue, has became a serious social problem. Therefore, seeking an on-vehicle, real-time and objective driving fatigue detection technology has a great significance for reducing the traffic accidents, safeguarding the security of the people's lives and property and enhancing the management of the passenger transport.In this thesis, an overview of a variety of driving fatigue detection method is introduced, and the PERCLOS method, which could utilize simple equipment, have low cost and could be more comfortable and reliable, is established as the research direction based on the face image processing technology.The main tasks of this thesis include:1. Expounding the Viola classic face detection method, which is a complete system comprised of haar-like feature, illumination balance preprocessing, cascade AdaBoost classifier and multi-scale problem solutions.2. Expounding the way to transplant the Viola method into the steps of driving fatigue detection, such as eye detection, eye state recognition and PERCLOS calculation.3. Utilizing OpenCV to realize the demo program about the driving fatigue detection comprised of the video image capture from the digital camera, the classifier training and the object detection.4. Exploring the LBP feature experimental application. Two methods are proposed, one is to scale the LBP feature, and another is to locate the eyes and recognize their states efficiently.5. Making the conclusion and the prospect of the full text.