A Network Optimized Ridge Estimator for Robust PSI Parameter Estimation and Its Application on Deformation Monitoring of Urban Area

Abstract

Persistent scatterer interferometry (PSI) has the ability to achieve submeter-scale digital elevation model (DEM) and millimeter-scale deformation. However, it's difficult to maintain good performance under long baseline and large phase gradient. In this paper, we develop a network optimized ridge estimator (NORM) for two-tier network PSI. Such a method can increase parameter estimation robustness and spatial continuity in single-prime mode and urban build environments. Firstly, we construct an optimized first-tier network by outlier detection and subsets network construction. On this basis, we utilize the ridge estimator as a substitute for the WLS estimator to improve the estimation accuracy in network adjustment. In the second-tier network, we detect the remaining PSs by constructing local star networks, thus increasing the density of the measurements in the first-tier network. The effectiveness of the proposed method is demonstrated by using the 2016-2018 TerraSAR-X dataset acquired in Shenzhen Metropolis, China. In addition, the stress deformation of two typical buildings, i.e., Shenzhen CITIC Building and Shenzhen Convention and Exhibition Center, are monitored and analyzed.