|
|||||
|
|
| 基于道路环境上下文的行人跟踪方法 | |
| 引用本文： | 方义,嵇智源,盛浩.基于道路环境上下文的行人跟踪方法[J].计算机应用,2015,35(8):2311-2315. |
| 作者姓名： | 方义 嵇智源 盛浩 |
| 作者单位： | 1. 中航发动机控股有限公司 科技与信息化部, 北京 101304; 2. 中华人民共和国科学技术部 高技术研究发展中心, 北京 100044; 3. 北京航空航天大学 计算机学院, 北京 100191 |
| 基金项目： | 国家自然科学基金资助项目(61472019);国家863计划项目(2013AA01A603)。 |
| 摘 要： | 针对目前城市交通中人车混行场景中行人跟踪效果不佳的问题,提出了一种基于道路环境上下文的行人跟踪方法。首先通过对道路环境上下文进行分析,建立道路模型;其次在道路模型的约束下建立行人与环境的交互运动模型;最后利用该模型进行行人的跟踪。在真实场景中的实验表明使用了道路上下文信息的跟踪方法与连续离散连续能量最小化的多行人跟踪方法相比,多目标跟踪准确度从47.6%提升至63.2%,多目标跟踪精度从68.8%提升至74.3%。数值结果表明道路上下文信息对于提高人车混行场景中行人跟踪效果的有效性。 |
| 关 键 词： | 多行人跟踪 人车混行 环境上下文 道路模型 环境交互 |
| 收稿时间： | 2015-01-12 |
| 修稿时间： | 2015-04-24 |
Pedestrian tracking method based on road environment context |
|
| FANG Yi,JI Zhiyuan,SHENG Hao.Pedestrian tracking method based on road environment context[J].journal of Computer Applications,2015,35(8):2311-2315. | |
| Authors: | FANG Yi JI Zhiyuan SHENG Hao |
| Affiliation: | 1. Department of Technology and Information, Aviation Industry Corporation of China, Beijing 101304, China; 2. High Technology Research and Development Center, Ministry of Science and Technology of the People's Republic of China, Beijing 100044, China; 3. School of Computer Science and Engineering, Beihang University, Beijing 100191, China |
| Abstract: | Since tracking pedestrians with car driving is still a problem in multi-object tracking, a pedestrian tracking method based on road environment context was proposed. Firstly, with the analysis of road environment context, an interaction algorithm based on the path model was proposed for pedestrians' motion prediction, then the model of pedestrians pedestrians mixed with vehicles was introduced. Finally, the model was applied in pedestrian tracking. The experimental results show that compared with discrete-continuous tracking approach, Multiple Object Tracking Accuracy (MOTA) of the proposed algorithm grows from 47.6% to 63.2% and Multiple Object Tracking Precision (MOTP) grows from 68.8% to 74.3%. The results prove the effectiveness of road environment context to improve the pedestrian tracking effect in mixed vehicle scene. |
| Keywords: | multiple pedestrians tracking pedestrians mixed with vehicles environment context path model interaction with environment |
| 本文献已被 万方数据 等数据库收录！ | |
| 点击此处可从《计算机应用》浏览原始摘要信息 | |
| 点击此处可从《计算机应用》下载全文 | |
