Detection of Land Subsidence using DInSAR in Urban Area from 2014 to 2019

Hello! I am currently working on my undergraduate research with the given title above.

I just want to ask the following:

  1. how can we overcome temporal decorrelation and atmospheric error in DInSAR?
  2. how to select images to be downloaded and how long should the interval be?
  3. How to select image pairs? which ones should be the master and slave image
  4. what is the workflow in generating vertical displacement using DInSAR?

Your response will be very helpful. Thank you.

1 Like

Have you seen this tutorial?
Deformation mapping with Sentinel-1

It contains information on the selection of images and discusses potential challenges.

Decorrelation is a physical progress, you cannot avoid or reduce it. The only chance we have for InSAR is the selection of suitable images.

The impact of atmosphere can be reduced by combining multiple displacement products (assuming that the deformation rate is constant over time), as suggested in the last part of the tutorial.

Some typical InSAR errors are also discussed at the end of this tutorial: DEM creation with Sentinel-1 - workflow and challenges

Hello ABraun, I also have the mission to carry out the posted topic, to comment a little I am working on measuring the subsidence of the land in an urban area this objective was previously achieved in a thesis using PSI with sentinel 1 images for a period of 1 year With software paid by someone else, my mission is to update that data for management 2019 and the present. I have searched for information on the subject, I also read thesis, this forum and the documents it recommended above.
I want to start with tests following a thesis that explains the steps (preprocessing with SNAP and then with STAMPS), having a couple of images the master would be the oldest and the other the slave, in a series of images the appropriate master image should be selected using InSAR Stack Overview (coregistered stack product to be selected), in this tool it requests as input a stack of co-registered images, I thought to use S1 TOPS coregistration where two images are used, my query is having 36 images to analyze to perform insar stack and make the S1 TOPS preliminary co-registration how should I proceed? Perform co-registration by pairs selecting the oldest image with the following (1 & 2, 1 & 3, 1 & 4 etc) or in chain (1 & 2, 2 & 3, 3 & 4), please collaborate with me.

Have you seen this topic? It summarizes the current workflows to follow for the PSI with Sentinel-1. StaMPS - Detailled instructions

But I don’t think the oldest image is suitable as the master. I would rather recommend one which lies in the middle (temporally) of the observed period to avoid too much decorrelation.

Hello. I am writing the latest dissertations on the subject of land subsidence using the Snap2stamps package and the psInsar method. Can anyone help me? Where can I find it?

you can find publications using snap2stamps with GoogleScholar

Thank you dear Abraun
Do you have any suggestions for accessing theses with the same subject and method?

the best way is a systematic search which combines keywords in the advanced search. Here you can define if a term should be inside the title (e.g. subsidence) or only in the text in general (e.g. snap2stamps).