计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (23): 12-13.DOI: 10.3778/j.issn.1002-8331.2008.23.004

• 博士论坛 • 上一篇    下一篇

基于非线性扩散的图像分解中的自动阈值估计

黄英君,张 军   

  1. 国防科技大学 信息系统与管理学院 系统工程系,长沙 410073
  • 收稿日期:2008-03-25 修回日期:2008-05-08 出版日期:2008-08-11 发布日期:2008-08-11
  • 通讯作者: 黄英君

Threshold estimating in nonlinear diffusions based SAR image decomposition

HUANG Ying-jun,ZHANG Jun   

  1. College of Information Systems and Management,National University of Defense Technology,Changsha 410073,China
  • Received:2008-03-25 Revised:2008-05-08 Online:2008-08-11 Published:2008-08-11
  • Contact: HUANG Ying-jun

摘要: 在对SAR图像进行处理如目标提取和图像压缩时,可以将图像分解为多个分量,这些分量代表了不同的图像特征,针对不同分量的特点分别选取最合适的处理方式,则能够得到性能上的极大提升。使用非线性扩散方法,将图像分解为表征大尺度结构信息的结构分量和表征小尺度细节信息的纹理分量,并对扩散过程中的阈值自动估计问题进行了研究,将拉依达准则引入基于非线性扩散的图像分解的自动阈值求解,并对多种扩散参数形式的扩散过程进行了实验和评估。

关键词: 图像分解, 非线性扩散, 拉依达准则

Abstract: With the processing of the SAR image such as targets recognizing and image compressing,the image can be decomposing into different components.Each components corresponds different properties,which can be processed by corresponding algorithm in order to get better performance.A nonlinear diffusions based image decomposing algorithm is proposed in this article,and a framework of 3σ criterion based threshold estimating is proposed.At last,the thresholds of three diffusion parameter types are estimated in tests.

Key words: image decomposing, nonlinear diffusions, 3σ criterion