嵌入式驾驶员状态检测算法的实现与优化  

Realization and Optimization of Embedded Driver Status Detection System

作　　者：张旭[1,2,3] 李亚利[1,2,3] 陈晨[4] 王生进[1,2,3] 丁晓青[1,2,3] 

机构地区：[1]智能技术与系统国家重点实验室,中国北京100084 [2]清华信息科学与技术国家实验室,中国北京100084 [3]清华大学电子工程系,中国北京100084 [4]宾夕法尼亚大学计算机与信息科学系,美国宾夕法尼亚州费城PA 19104-6303

出　　处：《自动化学报》2012年第12期2014-2022,共9页Acta Automatica Sinica

基　　金：国家高技术研究发展计划(863计划)(2009AA11Z214);国家自然科学基金(61071135);教育部博士点基金项目(20090002110077)资助~~

摘　　要：提出了一种可以在嵌入式平台上实时运行的驾驶员状态检测算法.状态检测采用了基于统计学习的Adaboost算法与动态建模算法.与传统的采用主动红外光的方法相比,本系统采用对人眼更为安全的被动式方法,且对光线的变化有更好的鲁棒性.算法的主要创新点是:1)提出了检测区域自适应调整的单双眼检测相结合的Adaboost人眼检测算法,提高了人眼检测的准确性与速度;2)提出基于高斯混合模型的人眼动态建模跟踪算法,自动提取驾驶员眼睛区域灰度分布的信息,实现了对不同驾驶员人眼的建模与跟踪定位.在多个公共数据集以及实车采集的视频上进行的实验表明,该算法能够准确判断驾驶员的状态,满足实时处理的要求.This paper presents a novel driver status detection algorithm, which can be processed in embedded platform in real-time. The proposed method is based on Adaboost and dynamic modeling algorithm. Compared to the traditional active infrared radiation method, our system employs a safer passive way and the algorithm is more robust to the various illuminations. There are two main contributions： 1） it mixes up the single eye and eye pairs detectors together and presents an adaptive detection region Adaboost eye detection algorithm, improving the detection rate and speeding up the eye detection process; 2） it presents a dynamic eye modeling tracking algorithm which is based on the Gaussian mixture model. The tracking algorithm can automatically extract image intensity distribution of the driver＇s eye region, thus can model and track eyes of different drivers. Experiments on several public databases and driving sequences taken in real car show that the proposed method can detect the driver status precisely and satisfy the real time processing requirements. Key words Embedded system, driver status detection, dynamic eye model, eye tracking

关 键 词：嵌入式系统 驾驶员状态检测 人眼动态模型 人眼跟踪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]