InSAR技术是研究地质灾害、地球物理等领域地表缓慢形变特征的重要手段。SAR卫星长期积累的影像，数据量巨大，导致解算效率降低；时间基线拉长，失相干效应严重。申请项目以Sentinel-1卫星作为研究对象，构建增强谱分集多视处理和参数优化估计模型，克服失相干效应对影像配准精度的影响；研究时序影像降维理论和最优分组方法，提出基于空间自适应滤波和Phase-Linking技术的影像相位补偿策略，借鉴高光谱影像降维技术实现复数影像的降维重构，揭示降维处理、形变监测精度与处理效率之间的关系，为降低影像信息冗余度，提高时序处理效率和相干目标空间密度提供理论依据。

InSAR has been demonstrated to be an important tool to analyze the slow surface movement in the fields of geological hazards and geophysics. However, the long-term SAR observations contain large amounts of data, slowing down the computation efficiency. The long temporal baselines cause decorrelation, limiting the applicability of InSAR technique. The proposed project aims to use Sentinel-1 dataset as the basis and construct constructs the enhanced spectral diversity multi-look processing and the parameter optimization estimation model to overcome the impact of incoherent effect on image registration accuracy; study the theory of time-series image grouping dimension reduction and the optimal grouping method and an image phase compensation strategy based on spatial adaptive filtering and Phase-Linking technology will be proposed. The dimension reduction and reconstruction of complex image is realized by using hyperspectral image dimension reduction technology. The optimal relationship between the dimension reduction processing, deformation accuracy and processing efficiency will be revealed, so that providing a theoretical basis for eliminating the redundancy of image information and improving the time-series processing efficiency and coherent target spatial density.