DEM Generation Issue


I am trying to make a series of DEMs for the area surrounding the Kalgoorlie open pit mine in Australia. However, when I create my DEM, I get some values that are almost 500 meters greater than values seen in the SRTM 3sec DEM. Additionally, I do not see relief patterns that I expected to see in the area. I have attached screenshots to better show my issue.

To reference, I am creating these DEMs by:

  1. S-1 TOPS Coregisteration
  2. Interferogram Formation
  3. TOPSAR-Deburst
  4. Multilooking
  5. Filtering
  6. Export to Snaphu
  7. Unwrap phase using Snaphu
  8. Snaphu Import
  9. Phase to Elevation
  10. Range doppler correction, selecting to include the sampling DEM (SRTM 3sec)

What is causing my generated DEM to differ so drastically from the SRTM DEM?

Thanks in advance for any help provided!

please have a look at these comments on interferometric quality of Sentinel-1 products:

Good materials on error sources of InSAR products:

What is the temporal and perpendicular baseline of your image pair? The only thing you can do to improve the interferogram quality is to select suitable images: Help in DEM Generation using Sentinel-1

There are some excellent documents on the selection of suitable images for interferometry by ESA: Source Section B- 1.3 Selecting images for InSAR DEM generation.

1 Like

Thank you for your feedback! I tried to use some of the helpful tips presented to others in your links, with no success. I suppose to the best for me to do is chose a more suitable satellite to use for DEM generation.

Thanks again!

where is your study area located? Does it contain much vegetation cover?

My study area is of south western Australia, around Kalgoorlie. There is vegetation, but it is rather minimal due to the arid environment.

actually not the worst conditions for DEM generation.

This pair could be suitable: temporal baseline 12 days and perpendicular baseline of 150m

maybe even this: temporal baseline 36 days and perpendicular baseline of 181m

Have you considered ALOS data? There are some image pairs from ASF

As ERS and ENVISAT are now also freely availably from ESA, you can consider them as well, especially because there was a period when ERS-1 and ERS-2 operated in an offset of 1 day. ESA offers to search for pairs according to their baseline configuration:

1 Like

I’ve tried a few image pairs, but I’ll try those pairs as well. And yes, I was considering switching to ALOS, but I’ll attempt to use the pairs you’ve suggested first. Thanks for all your help!

ALOS is quite tricky in SNAP because the orbit files are not always handled correctly which leads to strange interferograms: Problems producing interferograms with ALOS

Thanks for the heads up. I’ll be sure to watch for that! Thanks again for all the insight.