作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (4): 1-3. doi: 10.3969/j.issn.1000-3428.2011.04.001

• 博士论文 •    下一篇

智能监控系统中的运动目标检测算法

田 峥,徐 成,杨志邦,冯 堃   

  1. (湖南大学计算机与通信学院,长沙 410082)
  • 出版日期:2011-02-20 发布日期:2011-02-17
  • 作者简介:田 峥(1983-),男,博士研究生,主研方向:模式识别,计算机视觉;徐 成,教授、博士;杨志邦,博士研究生;冯 堃,硕士研究生
  • 基金资助:
    国家“863”计划基金资助项目(2007AA01Z104)

Moving Objects Detection Algorithm in Intelligent Surveillance System

TIAN Zheng, XU Cheng, YANG Zhi-bang, FENG Kun   

  1. (School of Computer and Communication, Hunan University, Changsha, 410082, China)
  • Online:2011-02-20 Published:2011-02-17

摘要: 为解决现有视频监控系统中目标检测算法无法应付复杂的室外环境且计算量和存储量较大等问题,将像素从RGB空间转换到YUV空间建立基于码本的背景模型,并单独对每个码字中的亮度分量进行高斯建模,提取运动目标的轮廓后,用连通区域算法对图像进行形态学处理。典型测试序列和ROC数据的对比实验结果证明该算法是高效和实用的,且易于在DSP或FPGA等嵌入式系统上实时实现。

关键词: 运动目标检测, 智能监控系统, 码本模型, 高斯混合模型, 嵌入式系统

Abstract: In order to solve the problems that the existing motion detection algorithm in surveillance system can not work well in complex outdoor scene, and need too much computation and memory, this paper proposes a moving objects detection algorithm based on improved codebook model. Pixels are converted from RGB space to YUV space to build the Codebook Model(CBM), and then the luminance component of each codeword is modeled by Gaussian model. The image is morphological processed by connected components algorithm. A test with the typical video sequences and the analysis of ROC data prove that the algorithm is effective and practical, and it can easily be implemented in embedded system such as DSP and FPGA.

Key words: moving objects detection, intelligent surveillance system, Codebook Model(CBM), GMM, embedded system

中图分类号: