Drowsy driving detection system

Drowsy driving in Turkey as in every country causes of great financial lossy and fatal traffic accidents. These systems which have aspects of the social and commercial value worth doing research on the issue. In this study, a Drowsy Driving Detection System is presented. The system is implemented with video-based test data. Application of the model is carried out using version 2.4.5 of OpenCV library. Visual C + + programming language was used in the Visual Studio 2010 environment. To test the system ZJU (Zhejiang University) database were used. This is a ort he database and that have been used on Drowsy Driving Detection System previously. Current Drowsy Driving Detection System mainly consists of 4 modules. The first module is detected face image from the video capture device. The second module is recognized person from positioned face images. The third module is located the eyes from positioned face images. The last module calculates the ratio of the eye blink from images of eyes and it decides whether to alert the driver by blink rate and regretting the driver looked at the road. In this study, the cascade classifier was used ort he detection of the face and eyes. For face recognition LDA (Linear Discriminant Analysis) method was used. ort he detection of eye state the method of LBP (Local Binary Pattern) is used which in recent years, pull interest with simplicity of operation and durability preferred for real-time applications. Due to, the proposed system?s testing database which the absence of light control and to the dynamic nature of the background and the same subjects without glasses and wore two kinds of glasses, the results of the the system obtained can be said to be at least partly close to actual data.