首页 | 本学科首页   官方微博 | 高级检索  
     

一种基于相关邻域模型的SAR图像迭代滤波方法
引用本文:郦苏丹,张翠,王正志,李广侠.一种基于相关邻域模型的SAR图像迭代滤波方法[J].数据采集与处理,2003,18(2):179-184.
作者姓名:郦苏丹  张翠  王正志  李广侠
作者单位:1. 国防科技大学自动控制系,长沙,410073
2. 解放军理工大学通信工程学院,南京,210007
摘    要:提出一种基于相关邻域模型的SAR图像迭代滤波方法。该方法引入相关邻域模型描述图像的局部特性。首先建立一系列反映方向信息的模板,每一个模板代表一种邻域结构。在SAR图像的统计性质基础上。对于每个邻域结构,利用Bayes方法。根据图像的观测值,获得强度期望值的最大后验概率(MAP)估计。运用模拟退火算法,根据每个结构的概率,随机选取一个邻域结构作为最适合结构,它所对应的估计值作为期望强度的估计。温度逐步趋向于O,该算法可以获得估计的全局最优解。

关 键 词:合成孔径雷达  图像迭代滤波方法  SAR  邻域模型  图像处理
文章编号:1004-9037(2003)02-0179-06
修稿时间:2002年7月30日

SAR Image Iterative Filter Method Based on Correlated Neighbourhood Model
LI Su-dan ,ZHANG Cui ,WANG Zheng-zhi ,LI Guang-xia.SAR Image Iterative Filter Method Based on Correlated Neighbourhood Model[J].Journal of Data Acquisition & Processing,2003,18(2):179-184.
Authors:LI Su-dan  ZHANG Cui  WANG Zheng-zhi  LI Guang-xia
Affiliation:LI Su-dan 1,ZHANG Cui 1,WANG Zheng-zhi 1,LI Guang-xia 2
Abstract:A kind of SAR image iterative filtering methods based on correlated neibourhood model is proposed. This method describes a local image property by a correlated neighbourhood model. Firstly, a series of templates reflecting direction information are established and every template represents a kind of neighbour structure. Then based on the statistical property of SAR image, by using Bayes formulation the MAP estimate of the real intensity by observation image values for every structure is obtained. Simulated annealing algorithm is used to get proper structure from the candidate structure according to their probability, the estimate value corresponding this structure is the estimate of the underlying cross-section. The global optimum result can be obtained by simulated annealing algorithm when the temperature is reduced to zero.
Keywords:synthetic aperture radar  filter  correlated neighbourhood model  simulated anneal
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号