作　　者： ; ; ; ; ; (李福樑）;

机构地区： 华南理工大学土木与交通学院

出　　处： 《交通信息与安全》 2014年第2期1-5,共5页

摘　　要： 随着智能交通的发展,道路行人主动安全的重要性愈发凸显,且越来越受到国内外研究机构的重视。选取Harr-like特征,利用积分图快速计算矩形特征,对传统的Adaboost算法引入优化更新权重的方法,提出1种基于级联分类器的行人检测方法,实现行人的实时区域快速检测。实验结果表明,基于改进Adaboost算法的快速行人检测与定位方法实时性强、准确率高,基于该算法的行人快速检测与定位系统可靠性高、鲁棒性强。 With the development of the intelligent transportation system （ITS） ,proactive pedestrian safety protec-tion has become increasingly significant and attracted the attention from many research institutions around the world .A fast pedestrian detection and location method based on the improved Adaboost Algorithm is introduced in this paper .The Haar-like features are selected and integral image is utilized to compute the features .An optimization method of updating the weights is put forward to improve the traditional Adaboost Algorithm .The experimental results show that the fast pedestrian detection and localization method has a better performance in accuracy and robustness .

关 键 词： 行人检测 算法 特征 级联分类器算法

领　　域： [交通运输工程] [交通运输工程]