基于红外图像的眼睛开闭检测方法

摘要

基于红外眼睛图像，提出一种红外环境下判断眼睛开闭状态的方法，并应用于疲劳驾驶检测当中。对图像采用基于余弦的Gamma校正压缩灰度值域的高低两端，同时扩展中间灰度值范围提高对比度，使用多尺度Retinex变换消除阴影的干扰以获取更准确的眼睛形态，根据眼角角度和上下眼睑间的高度在平面坐标上的分布特征对眼睛状态进行分类。实验结果表明，该方法可以有效地判断眼睛的状态。%Based on the infrared eye images,a method was proposed for recognizing the states of eyes in the infrared environ-ment.First,cosine based Gamma correction was adopted to adjust the range of gray value to enhance image contrast.After that, to eliminate the effect of shadow,multi-scale Retinex was utilized to obtain more accurate shape of eyes.Next,based on the open angle of canthus and the distribution characteristics of the distance between top and bottom eyelids on the two-dimensional coordinate plane,the states of eyes were classified.Experimental results show that the proposed method can detect the states of eyes with high accuracy.