An improved image-based iris-tracking for driver fatigue detection system

With rapid development of highway transportation and increasing number of vehicles, the incidence rate of road traffic accidents also sets record. Among them, fatigue driving has become the main reason of traffic accidents. In this paper, we utilize OpenCV (Open Source Computer Vision) to implement face detection as well as eyes detection on Android platform and to compare the frequency of eyes closed. A fatigue detection system which is able to determine the examinee's fatigue state accurately was established. Firstly, this paper improve the traditional face detection algorithm (Adaboost algorithm) and increase the efficiency of face detection. Secondly, we use the template matching method to track the eye (iris) in real time, constantly updating the iris template to improve the accuracy of matching. Finally, the implementation of system is described in detail under the Android platform. Experiments show that the system can make full use of real time nature and robustness of AdaBoost face detection algorithm, to achieve high detection rate.