Eye Tracing in Real-time Fatigue Driving Monitoring System

摘要

In the real-time fatigue monitoring system, accurate track of the driver's eyes decides the reliability and practicability of the system. In this paper the method we adopt is, for images collected by infrared CCD camera, the frame-difference method is primarily adopted to locate the area of head. Laterly, we judge the image type according to whether the statistic pixel is suitable to the characteristic of complexion. As for colored image, we make use of the characteristic of complexion clustering to locate the face then apply the hybrid projection function to track the eyes; as for the infrared image, we apply the image differential projection function to locate the candidated district of eyes, further locate the eyes through Hough Transform. The experiment results indicate that the proposed scheme can eliminate the impact brought by light changes in driver's cab and exactly trace the eyes; meanwhile it has applied values to some extent.