测绘学报 ›› 2015, Vol. 44 ›› Issue (3): 301-308.doi: 10.11947/j.AGCS.2015.20130782

• 摄影测量学与遥感 • 上一篇    下一篇

利用SAR影像配准偏移量提取地表形变的方法与误差分析

陈强, 罗容, 杨莹辉, 雍琦   

  1. 西南交通大学遥感信息工程系, 四川 成都 610031
  • 收稿日期:2013-12-19 修回日期:2014-11-25 出版日期:2015-03-20 发布日期:2015-04-01
  • 作者简介:陈强(1974—),男,教授,博士生导师,主要从事摄影测量与遥感的理论与应用研究. E-mail:swjtucq@sina.com
  • 基金资助:
    国家自然科学基金(41472255;51178404);四川省应急测绘与防灾减灾工程技术研究中心开放基金(K2014B006)

Method and Accuracy of Extracting Surface Deformation Field from SAR Image Coregistration

CHEN Qiang, LUO Rong, YANG Yinghui, YONG Qi   

  1. Department of Remote Sensing and Geoinformation Engineering, Southwest Jiaotong University, Chengdu 610031, China
  • Received:2013-12-19 Revised:2014-11-25 Online:2015-03-20 Published:2015-04-01
  • Supported by:
    The National Natural Science Foundation of China (Nos. 41472255;51178404);Open Research Fund by Sichuan Engineering Research Center for Emergency Mapping and Disaster Reduction(No.K2014B006)

摘要: 单一的InSAR观测技术可提取地表沿雷达视线方向(LOS)上的一维位移,而利用SAR影像配准过程中的同名像素偏移量可提取地表沿雷达方位向(近南北向)与距离向(近东西向)的二维形变场,与LOS方向的一维形变形成优势互补.本文在分析SAR影像配准偏移量提取地表形变场方法的基础上,推导建立了雷达方位向与距离向形变提取的误差模型,探讨了该方法提取地表形变的主要误差源.以Bam地震ASAR影像和玉树地震PALSAR影像为数据源,分别开展同震形变场的提取与误差分析试验,结果表明,基于SAR像素配准偏移量提取同震形变场的精度主要受匹配窗口尺寸与过采样因子影响,形变提取误差随匹配计算窗口的增大而减小,形变提取精度随过采样因子的增大有适量提高,地形起伏效应在高差较大的SAR影像区域中表现较为显著.

关键词: SAR影像配准, 像素偏移量, 同震形变场, 精度分析, 地形效应

Abstract: InSAR can extract one-dimensional deformation along the line of sight of radar. SAR image coregistration and pixel offsets can be used to extract two-dimensional deformation along the azimuth and range directions, and can be considered as the significant complement for InSAR deformation measurement. Analyzing the method of extracting deformation field with the SAR image coregistration and pixel offsets, this paper presents the error model of deformation extraction along the azimuth and range directions. The main error components for deformation extraction from the pixel offsets are addressed. The experiments of coseismic deformation field extraction and error analysis are carried out with ASAR images of Bam earthquake and PALSAR images of Yushu earthquake. The results show that the coseismic deformation extraction accuracy is significantly affected by matching window size and oversampling factor. The deformation extraction error decreases with the increase of matching window size, and deformation extraction accuracy can be slightly improved with the increase of oversampling factor. Terrain effect can cause significant pixel displacements in SAR images over high-relief areas.

Key words: SAR image coregistration, pixel offset, coseismic deformation, accuracy analysis, terrain effect

中图分类号: