Suppression of Coherence Matrix Bias for Phase Linking and Ambiguity Detection in MTInSAR

Extraction of Velocity Time Series With an Optimal Temporal Sampling From Displacement Observation Networks

“…The classical way consists of defining the weight as the inverse of the residual [14]. Here, a down-weight function inspired from [22] is used to be more robust to outliers.…”

“…where ψ, the Tukey's biweight function, which is a common down-weight function [22] robust to large outliers, is:…”

“…The uncertainty of the resulting time series can be obtained theoretically by propagation through the inversion. The uncertainty is assumed to be random and independent for each displacement, hence the probability theory lead to propagate the diagonal covariance matrix of the error of displacement observations [22], [24] as:…”

“…with n ef f the number of effective observations after outlier removal based on the biweight function [22].…”

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

“…Arc measurements that are contaminated by overwhelming outliers or gross errors should be discarded [23][25] [26]. The lowquality arcs can be detected by the residual estimates [31]. For each arc, the absolute parameters of PSs can be obtained by ( 4), and the residual becomes = res L BX δ (5) If the residual is large, it is likely that the absolute estimates on the PSs are unreliable and the corresponding arc should be discarded.…”