中文摘要：

本文利用长时间序列的相干目标分析方法研究地面沉降的形变规律．基于传统差分合成孔径雷达干涉测量方法，相干目标分析是利用时间序列多景影像获取地表稳定反射体——高相干点目标的新方法．实验区选择上海市主城区约100km^2区域，InSAR的实验数据采用覆盖1992～2000年间的25景ERS-1／2的单视复影像．在无先验沉降场模型的情况下，实验结果显示该实验区的地面沉降场在时间范围上跨越8年，平均观测时间采样率约为4个月，垂直形变精度优于5mm，满足我国对城市地面沉降形变观测的要求．研究结果表明该方法可以用于今后大时间／空间尺度上的高精度地表形变观测．

英文摘要：

Distribution of land subsidence could be obtained with long-term coherent targets analysis. Based on the conventional differential InSAR, coherent targets analysis is a new method to identify stable natural reflector, highly coherent pointy target, with multi-temporal SAR images. The test site in our experiment covers about 100 km^2 in the main downtown of Shanghai city. 25 scenes of single look complex images of ERS - 1/2 from 1992 to 2000 are chosen as the import data set. The result of CT points over 8 years with average temporal sampling of 4 months has been presented in this paper. Without any apriori deformation model, the accuracy of the land subsidence velocity from CTA is less than 5 mm, which satisfies the requirement in monitoring urban subsidence in China. This investigation shows the potential of CTA approach in monitoring land deformation on a large temporal/spatial scale.