互联网 qkzz.net
全刊杂志网:首页 > 女性 > 文章正文
刊社推荐

一种自适应阈值的运动目标提取算法


  摘 要:为了准确地划分运动目标和背景区域,提出一种自适应阈值的运动目标提取算法,对现有基于背景差的提取算法进行改进。本算法将运动目标和背景作为两个聚类,对图像中的点按像素灰度进行分类,以聚类间的方根—算术均值距离最大作为分割阈值选择的准则,使得运动目标提取算法中二值化阈值能够自动更新,从而实现对运动目标的准确完整提取。实验结果表明,该算法能够较准确快速地提取运动目标,并对环境亮度突变、背景存在微小运动等情况具有较好的鲁棒性。

  关键词:运动目标提取; 背景差; 阈值; 方根—算术均值距离

  中图分类号:TP391文献标志码:A

  文章编号:1001-3695(2010)06-2380-04

  doi:10.3969/j.issn.10013695.2010.06.110

  Adaptive threshold algorithm for extracting moving objects

  SUN Mingyu1, DING Ying1,2, LI Wenhui1, JIANG Qi1, XU Changqing1

  (1.College of Computer Science & Technology, Jilin University, Changchun 130012, China; 2. College of Computer Science & Technology, Changchun University of Science & Technology, Changchun 130022,China)

  Abstract:This paper presented a moving objects extraction algorithm based on adaptive threshold, which had improved extraction algorithm based on background subtraction. Moving objects and background were regarded as two clusters, and classified the intensity values at the pixels, then adopted the max clusters’ square root arithmetic mean divergence as the threshold selection rule. So the threshold could be automatic update, in order to accomplish exactly and perfectly extract. Simulation results indicate that motion objects can be extract correctly and fast by using the new algorithm, even in the case of dynamic natural environments since they include motions like swaying vegetation, breaking luminance etc.

......
很抱歉,暂无全文,若需要阅读全文或喜欢本刊物请联系《计算机应用研究》杂志社购买。
欢迎作者提供全文,请点击编辑
分享:
 

了解更多资讯,请关注“木兰百花园”
分享:
 
精彩图文


关键字
支持中国杂志产业发展,请购买、订阅纸质杂志,欢迎杂志社提供过刊、样刊及电子版。
关于我们 | 网站声明 | 刊社管理 | 网站地图 | 联系方式 | 中图分类法 | RSS 2.0订阅 | IP查询
全刊杂志赏析网 2017