| Driving safety is closely related to each family and each person, so how to improve driving safety has become a hot topic in today’s society. As one of the main causes leading to traffic accidents, fatigue driving poses a huge threat to drivers’ and pedestrians’ safety of life as well as their properties. So, how to prevent fatigue driving and minimize the losses which result from this has great practical significance for the protection of people’s lives and properties.There are three key features which are closing eyes, yawning, and nodding when a driver is fatigue. In order to determine whether the driver is in a state of fatigue driving, this dissertation mainly focuses on extracting these three features of drivers by image processing method. First of all, haar-like face detection algorithm is employed to locate the face area of a driver. This article introduces Kalman filter to track the movement of the driver’s face and predict the precise location of his/her face in the next frame, so as to reduce the searching area of face detection algorithm and improve the efficiency of the driver’s face detection. And, after a driver’s face is detected, three algorithms are proposed to extract three major fatigue features. The first one is to estimate the open or closed state of the driver’s eyes, and this algorithm finds an accurate sub-window which only contains a single eye, then a method to estimate the state of an eye is applied. The second one algorithm propose a novel approach which uses a ellipse to fit the contour of a driver’s month, and this approach is very easy and effective to get the parameters of yawning. Some trials are implemented to prove that this method can achieve expected results. The last one is to judge whether a driver is nodding by estimating the position of him/her. In addition, this dissertation designs a software system to implement and verify the availability and accuracy of these algorithms. And we have got expected experiment results. |
PDF Full Text Request