摘要
汽车司机疲劳驾驶是引发交通事故的一个重要原因,许多国家正积极进行疲劳报警器的研究。本文在比较各种疲劳的生理特征后采用PERCLOS,设计了一套红外采集系统减少环境的干扰,然后在红外序列图像中用差分法,圆频率滤波等方法提取脸部区域,定位眼睛,计算瞳孔的面积。实验结果显示该方法对光照以及司机脸部姿势的变化不敏感,定位准确,同时能够满足实时性要求。
Driving in case of drowse is a most important factor of traffic accidents. Now many countries are active in studying drowsy monitor. This article presents a method based on PERCLOS after compared with others . The article present a system fixed with infrared camera to reduce the influence of the surroundings, locate the face and eyes, and then calculate the area of pupils that are not being covered by the eyelids in the infrared images by means of difference method, circle frequency filter and others. It has been tested not sensitive to lights and head posture, running precisely in real time.
出处
《微计算机信息》
北大核心
2005年第08X期119-121,43,共4页
Control & Automation
