吉林大学学报(信息科学版)

• 论文 • 上一篇    下一篇

基于视频的运动目标跟踪算法

刘俊杰 1a , 杨 勇 1a , 才 华 1b , 曲福恒 1a , 李双鑫 2   

  1. 1. 长春理工大学 a. 计算机科学技术学院; b. 电子信息工程学院, 长春 130022; 2. 东北师范大学 图书馆, 长春 130024
  • 收稿日期:2016-08-30 出版日期:2017-05-25 发布日期:2017-06-07
  • 通讯作者: 杨勇(1970— ), 男, 长春人, 长春理工大学教授, 硕士生导师, 主要从事图形 图像处理研究, (Tel)86-13504323586(E-mail)yy@ cust. edu. cn。
  • 作者简介: 刘俊杰(1989— ), 男, 长春人, 长春理工大学助理实验师, 主要从事计算机图像处理研究, (Tel)86-13578807929 (E-mail)403192671@ qq. com。
  • 基金资助:
    吉林省科技发展计划基金资助项目(20130101179JC)

Algorithm of Moving Object Tracking Based on Video Images Sequence

LIU Junjie 1a , YANG Yong 1a , CAI Hua 1b , QU FuHeng 1a , LI Shuangxin 2   

  1. 1a. School of Computer Science and Technology; 1b. School of Electronic and Information Engineering,
    Changchun University of Science and Technology, Changchun 130022, China;
    2. Library Northeast Normal University, Changchun 130024, China
  • Received:2016-08-30 Online:2017-05-25 Published:2017-06-07

摘要: 为更加准确、 快速地检测与跟踪到运动目标, 将背景差分法和帧间差分法相融合对 CAMSHIFT
(Continuously Adaptive Mean-SHIFT)算法进行改进。 首先, 通过背景差分法和帧间差分法相融合确定目标所在
区域, 然后结合 CAMSHIFT 迭代算法实现目标跟踪。 实验结果表明, 该方法改变了传统 CAMSHIFT 算法需手
动选定目标和跟踪窗容易发散的局限性, 并提高了跟踪的准确性与稳定性。

关键词:  CAMSHIFT 算法, 帧间差分法, 目标跟踪, 背景差分法

Abstract:  Moving target tracking in video is an important research direction in the field of computer vision. In
order to detect and track the moving objects more accurately and quickly, we propose an improved CAMSHIFT
(Continuously Adaptive Mean-SHIFT) algorithm based on background subtraction and frame difference algorithm.
First of all, the target area is determined by the background difference method and the frame difference method,
then combining with the CAMSHIFT iterative algorithm the target tracking is realized. Experiments show that this
method has changed the traditional CAMSHIFT algorithm which needs manually selecting the target, has overcome
the limitations of diverging, and has improved the accuracy and stability of tracking.

Key words: target tracking, background subtraction, continuously sdaptive mean-SHIFT(CAMSHIFT), frame difference method

中图分类号: 

  • TP391. 41