In order to meet the vision in eye fatigue driving monitoring based on computer and real-time positioning accuracy needs, we present an approach of quick face detection in the background of fatigue driving. According to the character that the background in the video is almost fixed, the paper introduce the background subtraction and gray-level projection to detect the face quickly. Experiments demonstrated that our method is real-time, and robust under variable lighting conditions and various face orientations.%为了满足基于计算机视觉的疲劳驾驶监测中眼睛定位的实时性和准确性的需要,提出了疲劳驾驶视频监测背景下的快速人脸定位方法.依据疲劳驾驶监测中获取的视频背景相对固定的特点,将背景减除法引入到驾驶员人脸检测当中,与灰度投影法相结合实现快速人脸定位.实验表明此方法快速、有效,且对光照变化以及驾驶员人脸姿态的变化具有较好的鲁棒性.
展开▼
