计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (36): 187-188.DOI: 10.3778/j.issn.1002-8331.2008.36.053

• 图形、图像、模式识别 • 上一篇    下一篇

视频监控系统中阴影去除方法的研究

苑 颖,罗青山,曾贵华   

  1. 上海交通大学 电子工程系,上海 200240
  • 收稿日期:2008-05-30 修回日期:2008-08-22 出版日期:2008-12-21 发布日期:2008-12-21
  • 通讯作者: 苑 颖

Shadow removal in video surveillance

YUAN Ying,LUO Qing-shan,ZENG Gui-hua

  

  1. Department of Electronic Engineering,Shanghai Jiaotong University,Shanghai 200240,China
  • Received:2008-05-30 Revised:2008-08-22 Online:2008-12-21 Published:2008-12-21
  • Contact: YUAN Ying

摘要: 提出了一种基于颜色不变性和建立阴影高斯模型的阴影检测和消除的方法。首先对亮度小于背景的前景根据颜色特征的相近性进行划分,再利用阴影的光谱特性建立高斯模型,去除运动目标的投影。最后利用阴影的空间特性对图像进行后处理,完成阴影的检测与消除。实验结果表明,该方法能够有效地检测和消除阴影。

关键词: 阴影检测与消除, 颜色不变性, 阴影高斯模型, 相似区域的划分

Abstract: A new approach for shadow detection and elimination based on color invariants and Gaussian shadow model is presented in this paper.First authors extract and classify uniform-region,and then create a single Gaussian model for cast shadow,remove the cast shadow based on the Gaussian model and uniform-region.At last,this result is improved using spatial analysis.The results illustrate that the approach is efficient when handling widely different conditions that can occur.

Key words: shadow detection and elimination, color invariants, Gaussian shadow model, uniform-region classification