计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (13): 193-195.

• 图形、图像、模式识别 • 上一篇    下一篇

基于区域的烟尘图像分割方法

张晓梅1,王诚梅1,韩 琥2,3   

  1. 1.山东大学 计算机科学与技术学院,济南 250061
    2.中国科学院 计算所 智能信息处理重点实验室,北京 100080
    3.中国科学院 研究生院,北京 100049
  • 收稿日期:2007-11-08 修回日期:2008-01-31 出版日期:2008-05-01 发布日期:2008-05-01
  • 通讯作者: 张晓梅

Region based segmentation methods for soot image

ZHANG Xiao-mei1,WANG Cheng-mei1,HAN Hu2,3   

  1. 1.School of Computer Science and Technology,Shandong University,Ji’nan 250061,China
    2.Key Lab of Intelligent Information Processing,Institute of Computing Technology,CAS,Beijing 100080,China
    3.Graduate University of Chinese Academy of Sciences,Beijing 100049,China
  • Received:2007-11-08 Revised:2008-01-31 Online:2008-05-01 Published:2008-05-01
  • Contact: ZHANG Xiao-mei

摘要: 烟尘图像分割是图像分割方法的一个较新的应用方向。高效、准确的烟尘图像分割技术对于大气污染监控、森林火灾预警以及军事情报获取等方面都有重要的意义。将阈值分割、区域生长、区域分裂与合并等基于区域的方法用于烟尘图像分割,并对三种分割方法在烟尘图像分割中的性能表现进行了分析。研究结果对烟尘图像分割中算法的选型具有一定的参考价值。

关键词: 烟尘图像, 图像分割, 阈值分割, 区域生长, 区域分裂与合并

Abstract: The segmentation for soot image is a new application for soot image segmentation methods.It is important to segment soot images accurately and efficiently for surveillance of atmosphere pollution,forewarning of forest fire as well as acquirement of military information.In this paper,region based segmentation methods,such as threshold,region growing and region split and merge,are used to implement the segmentation for soot images.Also,the performance of three methods for soot image segmentation is analyzed and compared.The study can be useful for the selection of segmentation methods for soot image segmentation.

Key words: soot image, image segmentation, threshold segmentation, region growing, region split and merge