宇航学报 ›› 2012, Vol. 33 ›› Issue (8): 1112-1118.doi: 10.3873/j.issn.1000-1328.2012.08.016

• 电子信息 • 上一篇    下一篇

基于EM算法的InSAR区域网连接点检测

姜丽敏1,2,3, 陈曙暄4, 向茂生1   

  1. 1. 中国科学院电子学研究所微波成像技术国家级重点实验室,北京 100190;
    2. 中国科学院研究生院,北京 100190; 3. 北京机电工程研究所,北京 100074;
    4. 北京航天自动控制研究所, 北京 100854
  • 收稿日期:2011-09-05 修回日期:2011-12-15 出版日期:2012-08-15 发布日期:2012-08-23
  • 作者简介:10001328(2012)08111207
  • 基金资助:
    国家高技术研究发展计划(863计划)(2007AA120302)

Tie Point Detection in InSAR Block Based on EM Algorithm

JIANG Li   

  1. 1. National Key Laboratory of Science and Technology on Microwave Imaging, Institute of Electronic,
    Chinese Academy of Science,Beijing 100190,China;

    2. Graduate University of Chinese Academy of Sciences,Beijing 100190,China;

    3. Beijing Electrical and Mechanical Institute, Beijing 100074, China;
    4. Beijing Aerospace Automatic Control Institute, Beijing 100854,China

  • Received:2011-09-05 Revised:2011-12-15 Online:2012-08-15 Published:2012-08-23

摘要: 本文提出一种基于期望极大化(EM) 算法的干涉测量合成孔径雷达(InSAR)区域网连接点( TPs)检测新方法。利用混合模型将TPs检测问题归结为缺损数据下的最大后验概率( MAP)估计和最大似然(ML)估计,并将特征描述符的匹配概率融入到算法框架中,最后通过EM算法实现变换矩阵和点对应同时求解。同航带相邻SAR影像和相邻航带相邻SAR影像的实验验证了方法的有效性。

关键词: 期望极大化, InSAR区域网, 连接点, 混合模型, 极大后验概率, 最大似然

Abstract: A method for detecting Tie Points (TPs) in interferometric synthetic aperture radar (InSAR) block based on EM (Expectation Maximization) algorithm is proposed in this paper. In this method the problem of TP detection is cast into maximum a posterior probability (MAP) estimation and maximum likelihood (ML) estimation by using a mixture model under missing data. This statistical framework also takes into account matching probabilities of the feature descriptors attached to interest points, and then the transformation matrix and point correspondences are simultaneously solved by using the EM algorithm. Experiments on SAR images of both same strip and adjacent strips validate this method respectively.

Key words: EM, InSAR block, Tie points, Mixture model, MAP, ML

中图分类号: