A stack of 18 X-band SAR scenes from Capella Space was acquired over a section of Mexico City using the medium-inclination orbit satellite CAPELLA-C14. The images were acquired at approximately 3-day intervals between June 26 and August 15, 2024, covering a total of 47.2 days. The scenes have relatively short spatial baselines (mostly < 500 m, relative to a centered spatial reference), making the dataset well suited for interferometric analysis. Differential Interferometry and Persistent Scatterer Interferometry (PSI) were performed using the GAMMA Software (https://gamma-rs.ch). Thanks to the short time intervals and relatively short spatial baselines, high coherence differential interferograms were generated over this predominantly urban area. PSI was successfully performed using both single-reference and multi-reference approaches.
With the spatial and temporal baselines being uncorrelated, the dataset allowed the simultaneous estimation of deformation rates and point heights. Due to the relatively short observation period, the uncertainty in the deformation rate estimates was relatively high. After subtracting an estimate of the atmospheric phase (differential phase related to atmospheric path delays), the statistical estimation error of the deformation rate (in line of sight) was in the order of ~1 cm/year. Therefore, this time series is well suited for studying areas of rapid deformation, such as the subsidence observed in Mexico City. A longer time series would be required to reliably measure slower deformations. The statistical point height estimation, on the other hand, is quite accurate, with errors around 0.25 m rms, thanks to the relatively long baselines that are part of the stack.
As an additional application, an RGB composite image was created using coherence, mean backscatter, and temporal variability, effectively highlighting land cover types such as buildings, vegetation, and water bodies.
Read the entire Case Study here.