Driver's eye state recognition based on model

It is very significant to use machine vision to detect the driver's eye state and judge the driver's fatigue. In existing machine vision acquisition system, the eye state recognition is often detected with the characteristics of the iris. Aiming at the problems that the missing or changing of the iris information will lead to the false detection of the eye state, and the accuracy and real time capability can not be taken into account at the same time, based on analyzing the machine vision system and human eye state, considering both of the accuracy and real time capability of the recognition algorithm, the machine vision system imaging is used to better distinguish the non-iris feature of the eye state, an improved human eye state recognition method is proposed based on a fusion model. Furthermore, an embedded platform was built, and the algorithm identification comparison experiment was conducted. The results show that the fusion detection algorithm using the non-iris feature improves the recognition discrimination capacity of open eye and closed eye. Compared with the template matching method, the accuracy of eye state recognition is enhanced by 15%, and the time consumption is reduced by about 1/3. © 2016, Science Press. All right reserved.