基于机器视觉的驾驶疲劳检测方法

针对基于机器视觉技术的驾驶防瞌睡装置，讨论了当前的发展状况和应用情况，提出了基于红外光源、差分图像、Kalman滤波的系统方案。在红外光照射下，利用视网膜对不同波长红外光吸收率的显著差别，引起图像处理区域改变，同时利用神经网络辅助Kalman滤波器对眼部位置进行跟踪预测，实现司机在一定范围内活动时跟踪眼睛、测量眼睑和眼球状态的分析技术。该方法为检测司机在驾驶中是否困倦提供了关键的技术，实验表明该方法是有效的。

The Machine Vision- based Method to Detect Drowsy Driver

It introduces a machine vision based system used for detection of drowsy driver. Based on the summary of detection technologies, it gives the scheme integreated with infrared, difference image and Kalman filter. Because the infrareds of two wavelengths make two images in which the only difference is the intensity of the retinal reflections, and Kalman tracker is used to deal with the dynamics of head/eye movements, this method can track the eyes of the driver and analyze the states of the pupils. The system affords pivotal tech nology of the drowsy driver detection, and the experiments show it is effective.