日兰高铁巨野煤田段农田遍布，合成孔径雷达干涉（InSAR）的时间失相干严重，可用于时序InSAR（MT-InSAR）分析的永久散射体（PS）稀少。将SAR数据限制在失相干影响较弱的10月至次年4月初并联合PS和分布式散射体（DS）有望解决该问题。然而，受限于SAR卫星的重访周期，仅采用10月至次年4月初的SAR影像会导致数据量变少。而当SAR数据较少、相干性较低时，难以准确估计协方差矩阵和相干矩阵，使得现有的DS相位估计方法误差较大。为此，提出了一种基于Fisher信息量的DS相位优化估计算法，利用Fisher信息量调节各干涉对的权重，抑制低相干干涉对的影响。通过模拟数据和真实数据验证了算法的可靠性和可行性。另外，构建了联合PS和DS的小基线（SBAS）干涉处理框架，在增加观测方程的同时保证干涉对的相干质量，从而实现形变信息的稳健估计。利用2020年10月至2021年4月间的Sentinel-1 SAR数据获取了日兰高铁巨野煤田段地表沉降，并结合已有的监测资料分析了地表沉降的成因及时空演化信息。研究结果表明：采用上述方法，能够根据10月至次年4月初的少量SAR数据监测高铁沿线沉降情况；日兰高铁巨野煤田段沿线仍在持续沉降，3公里范围内的平均形变速率集中在-3.5 ~ -0.5 cm/yr之间，与2015至2019年的观测结果一致，未出现加剧现象；巨野煤田段存在可能由断层活化、深层地下水流失等因素间接造成的更大范围地表沉降，并且沉降靠近高铁侧，这一点需引起注意。

Farmland is distributed along the Ri-Lan high-speed railway,which may lead to a failure of interferometric synthetic aperture radar (InSAR) due to the temporal decorrelation.Besides,there is a lack of persistent scatterer(PS)for multi-temporal InSAR (MT-InSAR) analysis in the rural area with sparse buildings.It is a potential solution to these problems by using only the SAR data acquired from October to early April of next year to improve the coherence of interferogram and by incorporating the PS and distributed scatterer(DS)to increase the number of coherent point.Nevertheless,due to the limitation of imaging capability of spaceborne SAR system,the number of SAR images available over a period of about six months is often too small to accurately estimate the covariance and coherence matrix of DS(especially in the case of low coherent scenarios),which will degrade the performance of traditional methods.Fisher information index is a common measure for the information content that a random variable carries about an unknown parameter,thereby a new methodology based on the Fisher information index is presented to robustly estimate the DS phase.The new methodology is compared with the related traditional methods via simulation analysis and real data experiments with Sentinel-1 data.The results show that the presented methodology performs better.Besides,a SBAS baseline approach incorporating both PS and DS is presented to increase the number of observation equations,improve the coherence of interferograms and robustly estimate the ground deformation.The ground subsidence along Ri-Lan high-speed railway between October 2020 and April 2021 is extracted employing the presented approach by using C-band SAR acquisition constructed from two tracks of Sentinel-1 data over the region.Then,the characteristics and evolution of the ground subsidence is investigated by comparing with the previous data.The results show that the ground subsidence can be extracted based on a small SAR dataset acquired from October to early April in northern China by using the approach presented.The study also shows that the ground displacement rate ranges from -3.5 to -0.5 cm/a and agrees with the previous measurements without significant worsening.Besides,a secondary subsidence,probably caused by the groundwater outflow and fault instability due to mining,appears further away from the main subsidence basin and affects the regular operation of Ri-Lan high-speed railway,which needs to be further investigated.

1 DS相位估计方法

   1.1 CAESAR和EMI算法

   1.2 基于Fisher信息量的EMI算法：FEMI

2 实验验证

   2.1 模拟数据实验

   2.2 真实数据实验

3 联合PS和DS的SBAS处理框架

4 日兰高铁巨野煤田段地表形变

   4.1 实验区域概况及所用数据

   4.2 PSI与SBAS结果比较

   4.3 形变监测结果的可靠性与精度分析

   4.4 高铁沿线地表的形变分析

5 结论