摘要: |
马尔可夫随机场(MRF)可以很好地描述空间连续性,选择适当的邻域系统,能对图像的结构特征建模。利用以能量函数表示的联合概率分布,可以使用优化算法进行参数估计。高斯MRF能够准确、简洁地表示图像的纹理,而且具有线性特性,计算方便。本文回顾了在SAR图像处理中使用的MRF模型,详细说明了其中2种在图像复原及分割中的应用。 |
关键词: SAR图像处理 马尔可夫随机场 模拟退火 流域变换 四叉树分解 |
DOI:10.3969/j.issn.1001-893X. |
|
基金项目: |
|
Markov Random Field Models for SAR Images Processing |
|
() |
Abstract: |
By Selecting a proper neighborhood system and using the ability of Markov Random Field (MRF) to describe spatial dependence (continuity), MRF can be used to model the structural and textural behavior of images. Using the joint probability distribution in terms of an energy function, estimation of parameters can be performed by the stochastic relaxation algorithm. Gaussian MRF can represent a range of textures accurately and compactly and can be analysed tractably. In this paper,several MRF models are introduced, and the application of two models to restoration and segmentation of SAR images are presented in detail. |
Key words: SAR images processing,Markov Random Field(MRF),Simulated annealing,Watershed transformation,Quadtree decomposition, |