计算机与现代化

• 图像处理 • 上一篇    下一篇

一种基于改进ViBe的运动目标检测方法

  

  1. 广东工业大学信息与通信工程学院,广东广州510006
  • 收稿日期:2015-02-02 出版日期:2015-07-23 发布日期:2015-07-28
  • 作者简介:吴剑舞(1988-),男,福建莆田人,广东工业大学信息与通信工程学院硕士研究生,研究方向:嵌入式系统及智能家居; 翁玲瑜(1989-),女,江西上饶人,硕士研究生,研究方向:嵌入 式系统及智能家居。

An Improved Algorithm for Moving Object Detection Based on ViBe

  1. School of Information and Communication Engineering, Guangdong University of Technology, Guangzhou 510006, China
  • Received:2015-02-02 Online:2015-07-23 Published:2015-07-28

摘要:

摘要: 针对ViBe算法存在的不能有效去除背景噪声以及鬼影现象,提出一种结合帧差法和ViBe算法的改进算法。首先利用帧差法提取出运动目标的轮廓,然后根据得到的轮廓粗提取运动目标区域,最后
在得到的运动目标区域基础上分类进行ViBe算法更新背景并分割出完整的运动目标。实验结果表明,此算法能够有效抑制噪声以及鬼影造成的影响,在速度方面也比原算法有所提高。

关键词: ViBe算法, 帧差法, 运动目标检测

Abstract:

To solve the problem that traditional ViBe algorithm can not effectively remove the influence of background noise and ghost, an improved algorithm based on ViBe and
framedifference algorithm was proposed. Firstly, the framedifference algorithm was used to extract the contour of the moving object. Secondly, the contour was used to get
the general area of the moving object. Lastly, the object could be perfectly segmented from region after updating background using ViBe algorithm on classification based on the
general area. Experiments show that, this algorithm can suppress the influence of noises and ghost effectively, and its processing speed is better than the traditional ViBe
algorithm.

Key words: ViBe algorithm, framedifference algorithm, moving object detection

中图分类号: