测绘学报 ›› 2024, Vol. 53 ›› Issue (7): 1308-1320.doi: 10.11947/j.AGCS.2024.20230017

• 大地测量与导航 • 上一篇    下一篇

分布式散射体相位估计奇异值分解法

祝传广1(), 张继贤2,3(), 龙四春1, 杨容华1, 吴文豪1, 张立亚1   

  1. 1.湖南科技大学地球科学与空间信息工程学院,湖南 湘潭 411201
    2.莫干山地信实验室,浙江 湖州 313299
    3.国家基础地理信息中心,北京 100830
  • 收稿日期:2023-01-16 发布日期:2024-08-12
  • 通讯作者: 张继贤 E-mail:zhucg@hnust.edu.cn;zhangjx@casm.ac.cn
  • 作者简介:祝传广(1984—),男,博士,副教授,研究方向为InSAR理论与应用。E-mail:zhucg@hnust.edu.cn
  • 基金资助:
    湖南省自然科学基金(2023JJ30240);国家自然科学基金(41901373)

Phase estimation of distributed scatterer based on singular value decomposition

Chuanguang ZHU1(), Jixian ZHANG2,3(), Sichun LONG1, Ronghua YANG1, Wenhao WU1, Liya ZHANG1   

  1. 1.School of Earth Sciences and Spatial Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, China
    2.Moganshan Geospatial Information Laboratory, Huzhou 313299, China
    3.National Geomatics Center of China, Beijing 100830, China
  • Received:2023-01-16 Published:2024-08-12
  • Contact: Jixian ZHANG E-mail:zhucg@hnust.edu.cn;zhangjx@casm.ac.cn
  • About author:ZHU Chuanguang (1984—), male, PhD, associate professor, majors in theories and application of InSAR. E-mail: zhucg@hnust.edu.cn
  • Supported by:
    The Natural Science Foundation of Hunan Province(2023JJ30240);The National Natural Science Foundation of China(41901373)

摘要:

常规的分布式散射体(DS)相位估计方法需要生成全组合干涉对以构建样本协方差矩阵(SCM),然后根据SCM的统计特性估计DS相位,这一过程不但计算耗时,而且占据大量存储空间。本文提出了一种基于奇异值分解技术的DS相位快速估计方法(SVDI)。该方法分析的对象是单主影像干涉对组成的干涉相位矩阵而非全组合干涉对组成的SCM,因而可以有效提高计算效率、节省存储空间。并且,理论上证明了在一定条件下SVDI的结果与常规的特征值分解方法(EVD)是一致的。模拟数据和真实SAR数据的结果表明,SVDI算法有更高的计算效率,并且其相位估计精度以及形变解算精度与常规算法是一致的。

关键词: 分布式散射体, 相位估计, 样本协方差矩阵, 特征值分解, 奇异值分解

Abstract:

The covariance matrix is the basis for estimating the phase of distributed scatterer (DS) when using conventional algorithm. Therefore, a full combination of SAR data should be generated firstly to construct the sample covariance matrix (SCM). However, this process is not only computationally expensive but also consumes a large amount of storage space. In this paper, a fast algorithm, referred to as SVDI (SVD to interferometric phase matrix), for estimating the phase of DS based on singular value decomposition is proposed. SVDI estimates the phase of DS from the interferometric phase matrix constructed by single-master interferograms rather than the SCM constructed by multi-master interferograms (i.e., the full combination of SAR data). Therefore, SVDI can effectively improve the computationally efficient and save the storage space. Moreover, it is theoretically proved that the results of SVDI are consistent with the conventional eigenvalue decomposition (EVD) method based on an assumption. The simulated and real SAR data is used to verify the feasibility and reliability of SVDI. The experimental results show that the phase and deformation estimation accuracy of SVDI is consistent with that of the conventional method.

Key words: distributed scatterer, phase estimation, sample covariance matrix, eigen value decomposition, singular value decomposition

中图分类号: