华南理工大学学报(自然科学版) ›› 2009, Vol. 37 ›› Issue (1): 69-73.

• 电子、通信与自动控制 • 上一篇    下一篇

基于块区域统计矩的实时运动检测模型

吕常魁 王宁生 徐研   

  1. 南京航空航天大学 CIMS工程中心, 江苏 南京 210016
  • 收稿日期:2007-12-12 修回日期:2008-05-15 出版日期:2009-01-25 发布日期:2009-01-25
  • 通信作者: 吕常魁(1971-),男,博士,副教授,主要从事机器视觉及工业应用研究. E-mail:maillck@yahoo.com.cn
  • 作者简介:吕常魁(1971-),男,博士,副教授,主要从事机器视觉及工业应用研究.
  • 基金资助:

    国家“863”/CIMS重点应用示范工程项目(863-511-930-009)

Real-Time Motion Detection Model Using Block-Based Statistical Moment

Lü Chang-kui  Wang Ning-sheng  Xu Yan   

  1. CIMS Engineering Center, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, Jiangsu, China
  • Received:2007-12-12 Revised:2008-05-15 Online:2009-01-25 Published:2009-01-25
  • Contact: 吕常魁(1971-),男,博士,副教授,主要从事机器视觉及工业应用研究. E-mail:maillck@yahoo.com.cn
  • About author:吕常魁(1971-),男,博士,副教授,主要从事机器视觉及工业应用研究.
  • Supported by:

    国家“863”/CIMS重点应用示范工程项目(863-511-930-009)

摘要: 为了提高实时运动检测的鲁棒性,利用变异系数对图像序列帧的变化进行块区域统计矩描述,采用共线性检验算法获得变异系数的最大似然估计,然后构建了基于块区域变异系数最大似然比的三帧差分运动检测模型.实验结果表明,该模型能够较完整地检测到运动物体的轮廓边缘,且对于光照渐变、阴影及背景物体小幅度扰动等都具有较强的鲁棒性.

关键词: 运动检测, 颜色矩, 变异系数, 共线性检验

Abstract:

In order to improve the robustness of real-time motion detection systems, a variation coefficient is used to calculate the block-based statistical moment, to describe the variations of frame sequences, and the colinearity test is employed to obtain the maximum likelihood estimation of the variation coefficient. Then, based on the maximum likelihood ratio of the variation coefficient of successive blocks, a novel three-frame difference motion detection model is established. Experimental results indicate that the proposed model is suitable for the detection of whole silhouette of a moving object and is robust to the gradual illumination change, the shadow and the background vacillation, etc.

Key words: motion detection, color moment, variation coefficient, colinearity test