| 基于群体性行为模型的人群分析及其应用 |
| |
| 引用本文： | 陈威华,张旭,曹黎俊,黄凯奇.基于群体性行为模型的人群分析及其应用[J].微计算机应用,2014(1):20-25. |
| |
| 作者姓名： | 陈威华 张旭 曹黎俊 黄凯奇 |
| |
| 作者单位： | 中国科学院自动化研究所模式识别国家重点实验室＆智能感知与计算研究中心,北京100190 |
| |
| 基金项目： | 中国科学院战略性先导科技专项（XDA06030300）资助 |
| |
| 摘 要： | 基于视频分析的人群监控,涉及到获取人群行为和数量,这在智能监控领域具有重要的现实价值。本文建立基于运动特征的群体性行为模型,挖掘复杂人群场景中的群体行为,用于人群行为和数量的分析。群体性行为模型是一种主题模型（LDA）,通过样本学习,可以获得描述不同群体行为的特征集,用于人群分析。实验中,将群体性行为模型应用于挖掘监控场景下的不同人群行为及其特征集,并使用人工神经网络完成人数统计,统计正确率达到92.35%。
|
| 关 键 词： | 群体性行为模型 LDA 人群分析 人数统计 |
Crowd Analysis Based on Crowd Behavior Model and Its Application |
| |
| CHEN Weihua,ZHANG Xu,CAO Lijun,HUANG Kaiqi.Crowd Analysis Based on Crowd Behavior Model and Its Application[J].Microcomputer Applications,2014(1):20-25. |
| |
| Authors: | CHEN Weihua ZHANG Xu CAO Lijun HUANG Kaiqi |
| |
| Affiliation: | (Center for Research On Intelligent Perception and Computing( CRIPAC), National Laboratory of Pattern Recognition( NLPR), Institute of Automation, Chinese Academy of Science(CASIA), Beijing, 100190, China) |
| |
| Abstract: | Crowd surveillance based on video analysis is valuable in real scene in the field of Intelligent Visual Surveillance（ IVS）, which aims to analysis crowd behaviors and count the number of people in crowd. This paper presents a crowd behavior model based on motion features to recognize crowd behaviors in complex scenes and estimate the number of people in crowd. A Latent Dirichlet Allocation（ LDA） algorithm is used in this model. By training sample,the suitable feature sets for different crowd behaviors are used for crowd analysis. In experiments,the crowd behavior model is applied on calculating feature sets and analyzing different crowd behaviors. An artificial neural network（ ANN） is involved for people counting. And the accuracy of results reaches nearly 92. 35%. |
| |
| Keywords: | crowd behavior model LDA crowd analysis people counting |
| 本文献已被 维普 等数据库收录！ |
|
