摘要
虽然驾驶员的注视点定位对研究驾驶行为和提高驾驶安全有重要作用,但是光线不均匀和驾驶员头部的大角度转动造成了传统视线跟踪算法必需进行的校准过程在真实驾驶环境下无法实现.文中根据车辆环境和驾驶员注视行为特点,构建驾驶员注视视线区域和头部姿态的映射关系,以实现视线自动校准.文中利用稳健的面部特征定位及三维几何结构估计头部姿态,利用粒子滤波对头部姿态进行学习和跟踪.文中使用辅助采样的方法解决粒子滤波在给定的状态空间上的粒子贫化和权值分配问题.文中的自动校准算法误差较小,可以满足驾驶安全和驾驶行为研究的需要.
Driver＇s eye gaze tracking is very important for driver behavior study and driving safety.But non-uniform light and big rotation angle of head movement cannot make traditional eye gaze tracking algorithm be used in real driving environment.The proposed method based on vehicle environment and driver＇s eye gaze behavior characteristic is to implement automatic calibration by building the mapping between head pose and gaze zone.The head pose is estimated by robust facial landmarks detection and three-dimensional geometry configuration.And the head movement in the given state space is tracked and learned by particle filter.In order to solve the particle impoverish and weight assignation problem of particle filter,auxiliary sampling method is used.The experimental results show the error rate is low enough to meet the need of driver safety and driving behavior study.
出处
《计算机学报》
EI
CSCD
北大核心
2015年第12期2390-2399,共10页
Chinese Journal of Computers
基金
国家自然科学基金(61272368
61370142)
辽宁省高等学校优秀人才支持计划(LR2013021)
交通运输部应用基础研究项目(2015329225300)资助
关键词
视线跟踪
头部姿态
校准
粒子滤波
驾驶行为
eye gaze tracking
head pose
calibration
particle filter
driver behavior
