基于眼动参数云融合模型的疲劳检测方法  

Fatigue detection method based on eye movement-features fusion cloud model

作　　者：丁伍洋[1] 陈云华[1] 高素文[1] 张灵[1] 

机构地区：[1]广东工业大学计算机工程系,广东广州510006

出　　处：《计算机工程与设计》2013年第11期4067-4071,共5页Computer Engineering and Design

基　　金：广东省科技计划国际合作基金项目(2010B050400007)

摘　　要：针对采用确定性阈值法分析眼动特征,进而检测疲劳中存在的准确性差等问题,提出了一种基于不确定性云模型理论的眼动参数描述方法,实现了对眼动特征的量化、融合,以及最后疲劳状态的判定。结合Haar分类器与LBP算子定位检测眼部状态,提取眨眼时间和次数;采用云发生器算法建立了疲劳监测的眼动参数云融合模型,展开了相关实验研究。研究结果表明,与确定性阈值法相比,该方法检测的准确性更高,能作为驾驶员疲劳判别的可靠依据。For the poor accuracy using deterministic threshold to analyze the eye movement parameters in fatigue detection, a method based on the theory of uncertainty cloud model is proposed to describe the eye movement parameters, and the eye movement characteristics of quantification, fusion and fatigue judgment are realized. Firstly, eye state is located and detected combined with Haar classifier and LBP operator. And then the eye movement parameter fusion cloud model of fatigue monitoring is established using the cloud generator algorithm and experimental research is realized. The results show that this method achieve higher accuracy compared with the method using deterministic threshold, it can be used as a reliable basis for fatigue detection.

关 键 词：云融合 不确定性 LBP算子 眼动特征 疲劳检测 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]