基于快速区域定位和二重间隔分布机的行人检测  

Pedestrian Detection based on Fast Region Location and Double Distribution Support Vector Machine

作　　者：程凡永[1] 罗海波[1] 阮志强[1] CHENG Fanyong;LUO Haibo;RUAN Zhiqiang(Fujian Provincial Key Laboratory of Information Processing and Intelligent Control, Minjiang University, Fuzhou 350108, China)

机构地区：[1]闽江学院福建省信息处理与智能控制重点实验室,福州350108

出　　处：《湖南理工学院学报（自然科学版）》2017年第2期27-32,共6页Journal of Hunan Institute of Science and Technology(Natural Sciences)

基　　金：福建省重点实验室开放课题(MJUKF201737);闽江学院教改项目(MJW201142103;MJW201122064);闽江学院育苗项目(YKY13007);福建省科技厅项目(2017H0029);国家自然科学基金面上项目(61672216)

摘　　要：为实现快速和代价敏感性的行人检测,设计了基于BING目标检测和二重间隔分布机的行人检测框架.首先利用BING特征进行快速目标筛选,在此基础上提出了启发式的目标区域定位算法,通过目标选择和区域定位策略来获取行人所在的区域.区域定位减少了分类器的搜索空间,只需将二重间隔分布机对定位区域进行目标检测即可获得行人检测的结果,提高了行人检测的效率和速度.In order to quickly and cost-sensitively detect pedestrian,this paper designed pedestrian detection model based onBING and Double Distribution Support Vector Machine(DDM).Based on objectiveness proposals from BING feature,heuristicregion location algorithm was proposed.The region including pedestrian can be obtained by objectness select and region locationalgorithm.Region location effectively reduced the search space of DDM,therefore DDM only scan the location region to detectpedestrian to obtain detection result and position.Region location improves the efficiency and speed of pedestrian detection.

关 键 词：快速检测 区域定位 行人检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]