欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2011, Vol. 28 ›› Issue (2): 223-227.DOI: 10.7523/j.issn.2095-6134.2011.2.013

• 论文 • 上一篇    下一篇

一种用修正各向异性扩散方程抑制SAR相干斑的算法

张强1,2, 张毅1   

  1. 1. 中国科学院电子学研究所, 北京 100190;
    2. 中国科学院研究生院, 北京 100190
  • 收稿日期:2010-03-24 修回日期:2010-05-31 发布日期:2011-03-15

An improved anisotropic diffusion equation for SAR speckle reduction

ZHANG Qiang1,2, ZHANG Yi1   

  1. 1. Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China;
    2. Graduated University, Chinese Academy of Sciences, Beijing 100190, China
  • Received:2010-03-24 Revised:2010-05-31 Published:2011-03-15

摘要:

针对各向异性扩散方程抑制SAR图像相干斑时,需要进行多次迭代,很难用于实时处理的情况,提出一种改进算法.该算法根据图像变差系数,自适应调整中心像素与各方向扩散系数的权重,从而在均匀区域加速扩散过程,在非均匀区域保留原各向异性扩散方程的性能.利用机载SAR图像对算法性能进行验证,结果表明,算法仅需较少的迭代次数,就可以获得与常规各向异性扩散方程相干斑抑制相同的性能.

关键词: 合成孔径雷达图像, 相干斑, 滤波, 各向异性扩散方程

Abstract:

In using anisotropic diffusion equation to reduce SAR image speckles, a number of iterations are needed, which makes it impossible to deal with real-time processing. To solve this problem, an improved algorithm is presented in the paper. Weights of the central pixel and all the direction diffusion coefficients are adaptively adjusted according to image differential coefficients of variation. Diffusion process is accelerated in the homogeneous region, while performance remains in inhomogeneous region. Airborne SAR image was used to verify the effectiveness of the approach. Less iterations are needed in the improved method than in the normal method.

Key words: SAR image, speckle, filter, anisotropic diffusion equation

中图分类号: