计算机应用 ›› 2014, Vol. 34 ›› Issue (6): 1746-1752.DOI: 10.11772/j.issn.1001-9081.2014.06.1746

• 虚拟现实与数字媒体 • 上一篇    下一篇

基于隐马尔可夫模型的多摄像头人体对象的目标识别

高鹏1,郭立君1,朱一卫2,张荣1   

  1. 1. 宁波大学 信息科学与工程学院,浙江 宁波 315211;
    2. 国家电网浙江省电力公司 宁波供电公司,浙江 宁波 315099
  • 收稿日期:2013-12-04 修回日期:2014-01-28 出版日期:2014-06-01 发布日期:2014-07-02
  • 通讯作者: 高鹏
  • 作者简介:高鹏(1989-),男,湖北枣阳人,硕士研究生,主要研究方向:计算机视觉;郭立君(1970-),男,辽宁凌源人,副教授,博士,主要研究方向: 计算机视觉、模式识别;朱一卫(1970-),男,浙江诸暨人,工程师,主要研究方向:模式识别、电网优化;张荣(1970-),女,河南鹤壁人,副教授,博士研究生,主要研究方向:数字图像取证、模式识别。
  • 基金资助:

    重庆市科委基础与前沿研究项目;浙江省新一代移动互联网用户端软件科技创新团队项目;宁波大学人才工程项目;宁波大学人才工程项目

Multi-camera person identification based on hidden markov model

GAO Peng1,GUO Lijun1,ZHU Yiwei2,ZHANG Rong1   

  1. 1. College of Information Science and Engineering Ningbo University, Ningbo Zhejiang 315211, China;
    2. Ningbo Power Supply Company, State Grid Zhejiang Corporation, Ningbo Zhejiang 315099, China
  • Received:2013-12-04 Revised:2014-01-28 Online:2014-06-01 Published:2014-07-02
  • Contact: GAO Peng
  • Supported by:

    National Natural Science Foundation

摘要:

在非重叠多摄像机系统的人体对象目标识别中,针对基于单幅图片的识别算法不能较好处理对象表观和视角变化的问题,提出基于人体图像序列的算法。该算法用隐马尔可夫模型(HMM)融合多幅图片的特征,先考虑人体结构的约束,将人体图像在垂直方向上划分为多个相等的图像区域;然后采用多层阈值分割算法提取区域代表性颜色特征(SRC)和标准差特征(SSV);再用每个人体对象的多幅图片提取的特征数据集训练该对象的连续密度HMM;最后利用训练的模型实现人体对象的目标识别。该方法在两个公开数据集上进行的实验都获得了较高的识别率,提高了对摄像头视角变化、低分辨率的鲁棒性,且简单易实现。

Abstract:

In the non-overlapping filed of multi-camera system, the single-shot person identification methods cannot well deal with appearance and viewpoint changes. Based on the multiple frames acquired from surveillance cameras, a new technique which combined Hidden Markov Model (HMM) with appearance-based feature was proposed. First, considering the structural constraint of human body, the whole-body appearance of each individual was equally vertically divided into sub-images. Then multi-level threshold method was used to extract Segment Representative Color (SRC) and Segment Standard Variation (SSV) feature. The feature dataset acquired from multiple frames was applied to train continuous density HMM,and the final recognition was realized by these well-trained model. Extensive experiments on two public datasets show that the proposed method achieves high recognition rate, improves robustness against viewpoint changes and low resolution, and it is simple and easy to realize.

中图分类号: