Generation of Earth’s Surface Three-Dimensional (3-D) Displacement Time-Series by Multiple-Platform SAR Data

Abstract

In this chapter, the recent advancements of differential synthetic aperture radar interferometry (DInSAR) technique are presented, with the focus on the DInSAR-based approaches leading to the generation of three-dimensional time-series of Earth's surface deformation, based on the combination of multi-platform line-of-sight (LOS)-projected time-series of deformation. Use of pixel-offset (PO) measurements for the retrieval of North-South deformation components, which are difficult to be extracted from DInSAR data, only, is also discussed. A review of the principal techniques based on the exploitation of amplitude and phase signatures of sequences of SAR images will be first provided, by emphasizing the limitations and strength of each single approach. Then, the interest will be concentrated on the recently proposed multi-track InSAR combination algorithm, referred as minimum acceleration InSAR combination (MinA) approach. The algorithm assumes the availability of two (or more) sets of SAR images acquired from complementary tracks. SAR data are pre-processed through one of currently available multi-temporal DInSAR toolboxes, and the LOS-projected surface deformation timeseries are computed. An under-determined system of linear equations is then solved, based on imposing that the 3-D displacement time-series have minimum acceleration (MA). The presented results demonstrate the validity of the MinA algorithm.

Theory and Statistical Description of the Enhanced Multi-Temporal InSAR (E-MTInSAR) Noise-Filtering Algorithm

“…As evident, as the baseline threshold decreases and sets of SB interferograms are selected the number of coherent pixels increases. This testifies that the standard deviation of the original interferograms has a role in the optimization procedure, as indicated by the sample covariance matrix relationship expressed by Equations (50,51). Finally, Figure 14 plots the number of the whole coherent pixels of the set of SB interferograms corresponding to the test carried out by using a perpendicular baseline threshold of 300 m (which has been used as a reference).…”

“…This paper is, at most, focused on the presentation of the statistics fundaments of the core E-MTInSAR algorithm, instead of the general performance of the improved EMCF-SBAS processing chain. Interested readers can find additional details on the improved EMCF-SBAS InSAR toolbox, as well as an overview of some experimental results in the literature, see for instance [48][49][50]. Figure 15 shows a comparison between the average spatial coherence of the original (unfiltered) and the reconstructed (filtered) sequence of multi-looked interferograms for the selected case-study area.…”

“…Nonetheless, the identification of statistically homogenous pixels (SHPs) and the implementation of a suitable adaptive multi-looking step, as suggested in Reference [30], might be considered as a further evolution of the E-MTInSAR algorithm. This research study has demonstrated from a mathematical and statistical perspective that: (i) The phase estimator in Equation 12is consistent and the measured (optimal) weighted circular variance value gives a direct figure of the standard deviation of the circular variance estimator in Equation (12); (ii) the standard deviation of the circular variance estimator adopted in Equation 12critically depends not only on the number of used interferograms, namely M, but also on the absolute value of the perpendicular baseline of these interferograms, in particular, as the average absolute perpendicular baseline approaches the critical baseline the estimator diverges; (iii) the actual standard deviation of the reconstructed, noise-filtered interferograms strictly depends on the phase stability of the original interferograms, the overall graph network of used interferograms, see Equations (50) and (51), and the presence of the time-inconsistencies, namely the signal D, in the original interferograms. Equation (51) opens up the possibility to evaluate the global performance of the algorithm in terms of phase stability of the reconstructed interferograms.…”

“…The outcome expressed by Equations (50) and (51) is significant, even it is not definitive. This research paper represents a step forward to Reference [26].…”

Landslide hazard assessment using advanced remote sensing techniques and GIS

“…InSAR techniques can identify ground-surface deformations depending on the phase shift between the same ground pixels of at least two different SAR acquisitions from different passes for the same portion of terrain. Therefore, in the case of good coherence between two SAR images, it is possible to obtain information with sub-centimetre precision about ground-surface deformations in the satellite Line-Of-Sight (LOS) direction by unwrapping the displacement phase of the interferogram generated from these two images (Gens and Van Genderen, 1996;Fielding et al, 1998;Vasco et al, 2010;Ng et al, 2012;Capes and Teeuw, 2017;Pepe, 2017;Hammad et al, 2018b). The phase shift can be resulted as a set of interferometric fringes appear as cycles of colours in the displacement phase of the interferogram generated from the two SAR images.…”