期刊文献+

自动门限技术在靶纸图像目标分割中的应用

A Research on Automatic Threshold Techniques Applied to ObjectSegmentation of Shooting Target Image
下载PDF
导出
摘要 在图像自动报靶系统中,需要自动快速地进行靶纸图像中目标区域的分割,选取合适的自动门限分割方法是其关键。文中在分析靶纸图像特点的基础上,对目前两种性能较优的自动门限技术:类别方差自动门限法和矩不变自动门限法进行了介绍,并对两者的获取最佳门限值进行了比较。试验结果表明,类别方差自动门限法获取的门限值普遍偏大,文中简要分析了其不可避免的原因;而矩不变自动门限法由于根据的是图像自身的特征,获取的最佳门限值比较合适,更适用于靶纸图像中目标分割,且具有运算速度快的优点。 It is very important that selecting a right automatic segmentation method which can segment object regions of a shooting target image automatically in automation scoring system.Based on analyzed attributes of shooting target images, discuss two best automatic threshold techniques, one is Class Variance, the other is Moment Invariance, and compare their real threshold. A lot of experiments show that the threshold of Class Invariance is bigger than that of Moment Invariant. Simply analysis the procedure of Class Invariance method and find that it is inevitable to acquire a bigger threshold value, on the contrast , the threshold value acquired by the Moment Invariant method is more adaptive to real images according to its own attributes. In brief, Moment Invariant method has a more better segmentation and it runs more fast when it apply to object segmentation of shooing target images.
作者 张红民
出处 《微机发展》 2003年第9期8-10,共3页 Microcomputer Development
关键词 靶纸图像 目标分割 自动门限技术 图像分割 图像处理 边缘检测 模式识别 image segmentation automatic threshold techniques class variance moment invariant automatic scoring
  • 引文网络
  • 相关文献

参考文献4

  • 1张红民,何健鹰.图象中一种不规则边界圆的参数求取方法[J].计算机工程与应用,2002,38(16):75-76. 被引量:4
  • 2Pei S C,Homg J H. Circular Arc Detection Based On Hough Transform[ J ] .Pattern Recognitiaa Letter, 1995, (16) : 615 -625.
  • 3Sahoo P K, Soltani S, Wong A K C, et al . A Servey of Thresholding Techniques[J]. Comput Vision Graphics Image Process, 1988,41:233 - 260.
  • 4Lee S U, Chung S Y, Park R H. A Comparative Performance Study of Several Global Thresholding Techniques for Segmentation[J]. Comput Vision Graphics Image Process. 1990,52:171 - 190.

二级参考文献1

共引文献3

;
使用帮助 返回顶部