• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 论文 • 上一篇    下一篇

人数统计与人群密度估计技术研究现状与趋势

张君军,石志广,李吉成   

  1. (国防科技大学电子科学学院ATR实验室,湖南 长沙 410073)
  • 收稿日期:2016-03-29 修回日期:2016-11-04 出版日期:2018-02-25 发布日期:2018-02-25

Current researches and future perspectives of crowd
counting and crowd density estimation technology

ZHANG Jun-jun,SHI Zhi-guang,LI Ji-cheng   

  1. (ATR Laboratory,College of Electronic Science,
    National University of Defense Technology,Changsha 410073,China)
  • Received:2016-03-29 Revised:2016-11-04 Online:2018-02-25 Published:2018-02-25

摘要:

人数统计与人群密度估计是人群分析中的重要分支,也是视频监控所关注的重要信息之一。尽管近几十年来该领域取得了一些重要进展,但仍存在一些具有挑战性的问题。综述了基于计算机视觉的人数统计与人群密度估计方法的研究现状以及发展动态。首先,介绍了人数统计与人群密度估计技术的发展背景及应用方向。其次,总结了近年来提出的比较重要的方法,从机器学习的角度,将其分为浅层学习的方法和深度学习的方法;而从学习到的模型角度又可将其分为直接的方法(即基于检测的方法)和间接的方法(如基于像素的方法、基于纹理的方法以及基于角点的方法)。详细介绍了近二十年来基于浅层学习的方法,并对近些年来基于深度学习的人数统计与人群密度估计技术做了一个简要的总结。然后,对人数统计及人群密度估计方法性能评估技术进行简介,并提供了几个用于人数统计与人群密度估计的测试与评估数据集。最后,总结了该领域存在的技术挑战并对未来的研究方向进行了展望。

 

关键词: 人数统计, 人群密度估计, 浅层学习, 深度学习

Abstract:

Crowd counting and crowd density estimation are the important branch in crowd analysis, and are also the important information that surveillance always concerns. Although some important progress has been made in this field in recent decades, there are still some challenging problems. This paper reviews the current researches and development trends of the crowd counting and crowd density estimation methods based on computer vision. Firstly, the development background and application direction of the crowd counting and crowd density estimation technology are introduced. Secondly, the important methods proposed in recent years are summarized, which can be divided into two types from the machine learning point of view: the shallow learning based methods and the deep learning based methods. In the other hand, they can be divided into two types from learning model standpoint: the direct method (i.e., the detection based method) and the indirect method (i.e., the pixel-based, texture-based and corner point based methods). This paper introduces the shallow learning based methods in the last twenty years in detail, and makes a brief summary of the deep learning based methods in recent years. Then, a brief introduction is made on the performance evaluation techniques of crowd counting and crowd density estimation methods, and several data sets are provided for testing and evaluating these methods. Finally, the technical challenges in the field are summarized and the future research directions are prospected.
 

Key words: crowd counting, crowd density estimation, shallow learning;deep learning