#9 Using the right kind of PS-InSAR – Different Tools for surface motion estimation with SAR
Half-Day Tutorial for students, researchers, and practitioners working or interested in the field of PS-InSAR and related techniques. It is intended for beginners, but also experienced users struggling with the variety of available methods and the selection of the correct tools. Being a specific and advanced topic, this might be of interest for a comparably small number of participants, estimating around 20-30.
Synthetic Aperture Radar (SAR) systems are excellent tools for measuring range differences. This feature can be used for DEM generation using trigonometry. Furthermore, it provides a direct measurement for measuring motion. Using differential SAR interferometry (D-InSAR) slow motions can be detected with a precision in the millimeter scale. However, there are various error sources and limitations to the approach, making it basically unusable for many applications.
First of all, the interferometric measurement uses the phase difference between two acquisitions, where the phase is limited to values from – 𝜋 to 𝜋, limiting unambiguous measurements to this range. SAR is a coherent imaging system and is therefore suffering from speckling. Due to the long wavelengths, only a limited number of scatterers can be found per resolution cell, causing speckling that not only affects the appearance, but also renders the phase of a single resolution cell with distributed scattering nearly unusable. Changes on the ground between acquisitions lead to a loss of coherence between the signals. If these changes are too significant, the phases completely lose coherence causing the phase difference to be meaningless. Finally, atmospheric disturbances can have a significant effect on the phase difference. Although the general textbook wisdom is that SAR is independent from weather conditions, which is anyway only partly true for SAR amplitude images, it is certainly untrue for interferometric SAR. Measuring differences in path lengths from satellites in millimeter precision, clearly path delay differences through the atmosphere play a significant role.
In the last two decades, significant efforts have been undertaken to overcome the above mentioned problems. Starting with the PSInSAR approach developed around the year 2000, various additional methods have been developed that try to overcome these problems mainly by focusing on coherent targets and using a large number of SAR images to carefully separate
the deformation signal of the phase differences from the other signal contributions and error sources.
This approach was hugely successful and a variety of other approaches have been developed since, each adressing different issues and specializing on certain tasks and domains.
AUer 20 years of multi-baseline InSAR methods, the power of these techniques are well-known and with the increased data availability the demand for using the techniques has increased drastically since the launch of the commercial high-resolution satellites in 200ti, but even more with the global data availability of Sentinel-1. However, users may feel overwhelmed by the variety of the different approaches. This tutorial aims at overcoming the confusion by explaining the core concepts of multi-baseline InSAR and the main methods used.
Starting from a quick introduction into D-InSAR, the concept of the traditional PSInSAR approach is described. From this onward, different approaches, like StaMPS, STUN, SBAS, PSP, QPS, and SqueeSAR are introduced and the differences are explained. From the differences between them, the various applications are shown.
However, surface motion estimation is not limited to InSAR based methods alone. Pixel- tracking and Point-Target Offset Tracking will be introduced. Furthermore, SAR geodesy and TomoSAR will be introduced and the various methods will be put into context.
Course materials and resources
A set of lecture notes is distributed to the participants. We will prepare a set of training data for experiments with different approaches. We will offer hands-on with StaMPS and WhuPSI. Furthermore, we plan to include SarProZ hands-on experience, as it is allowing us to test the traditional PSInSAR approach as well as advanced methods, like QPS.