How to import an external DEM into SNAP?

Hello everyone,
I have created a mosaic of images from ALOS with a resolution of 12.5 m to calibrate may imagens of ALOS-2 (resolution of 10 m) and I don’t how to import the file into SNAP.
How can I do that? And if is possible to import it where can I access the file after import it?

Thank you very much!

Claudia Arantes

can you please speficy what you mean by “calibrate” the ALOS-2 images? Do you mean radiometric correction or geometric correction?

Normally, there there is no need to import external DEMs: Both the “Radiometric Terrain Flattening” and the “Range Doppler Terrain Correction” operators allow to select an external DEM, as long as it is stored in GeoTif format and projected to WGS84 (EPSG:4326).

Hello ABraun,

My ALOS-2 images are SLC, Stripmap, Fine. I am applying the following steps:
1 - Calibration to Beta 0
2 - Specke filtering
3 - Radiometric terrain flatening
4 - Range Doppler terrain correction
On steps 3 and 4 ask for a digital elevation model. And on this steps I’m trying to select a mosaic from Alos I have created. Are these steps correct?
Another question: Does it make much difference to use the SRTM-3sec or ALOS 12.5m that I had created is better?

Thank you very much,
Claudia

yes, these steps are correct and you can use external DEMs with the specifications I mentioned above (file type and projection).
If the spatial resolution is higher , the chance that (radiometric and geometric) topographic effects are corrected are clearly higher and you will have better SAR products.

But may I ask what you mean by ALOS 12.5 m and where you got it from? There is one larger misconception about this, but I’m not sure if it is the case here.

I’m doing the processing of SAR images with 10m of spatial resolution.
I did the download of DEM from ALOS images from VERTEX website (12.5m of spatial resolutuion). Is it not correct? Can I use then SRTM-3sec?

Unfortunately, this data not really offers a spatial resolution of 12.5 meter. For most parts of the world, the DEM contained in ALOS products fom Vertex is just SRTM data (30 m) which was up-sampled to a pixel size of 12.5 m. Please have a look here:

So, to answer your question: If you select SRTM 1Sec (AutoDownload), you already have the best quality which is freely available. I suggest to directly use it (instead of an external DEM, unless its quality is really better) to avoid unneccessary resampling.

1 Like

Thank you very much for your explanation!
I’ll use SRTM 1ArcSec.

Regards,

Claudia

ABraun,

In the Range Doppler Terrain Corrrection there is two options: SRTM 1sec grid and 1sec HGT. Which shoul I choose? My area is in Brazil (Amazonia Forest).
Thank you,

Claudia

1 Like

Both SRTM 1Sec are the same data source.

  • SRTM 1Sec HGT (AutoDownload): This is the one you select if you want SNAP to check and download the DEM data (recommended).
  • SRTM 1Sec Grid: You only select this option if your internet connection is not stable or your firewall prevents SNAP from accessing the online resources. In such a case you can download/copy SRTM data manually, but this only works if the folder structure and file names are correct (not recommended).

grafik

1 Like

Thank you very much!
Regards,
Claudia

Hello ABraun,
In the option Radiometric Terrain Flattening I select Beta 0 but the result is Gamma 0? Why?
Thanks,
Claudia

Beta0 is a lower level of calibration, taking into account the calibration constant. Gamma0 furthermore includes the local incidence angle. Some more explanations on this are here: Radiometric & Geometric Correction Workflow

Due to a firewall I have no control over I need to use SRTM 1 Sec Grid, but see that “this only works if the folder structure and file names are correct”. Can anyone please tell me the exact details regarding:

Which format of SRTM I should use (and from which website)?
What the “folder structure and file names” need to be in order for this to work?

Alternatively, is it just easier to download SRTM tiles, mosaic them (would any other pre-processing be required?) and use as an External DEM? I see any external DEM needs to be in WGS84 projection and saved as GeoTIFF, but are *.tfw and *.prj files also expected by SNAP (I’m able to output with or without such).

Many thanks in advance :slight_smile:

1 Like

This is the option which is less prone to errors, but it’s only applicable if your study area is constant, so you can prepare the SRTM mosaic once.
The GeoTiff without tfw and prj should work best as external DEM. No further processing needed besides projection and file format.

1 Like

Hi ABraun
Can you tell me what you mean by “projected to WGS84 (EPSG:4326)”?

Hi ABraun
In the second place , we can make firewall can 't interfere with work of SNAP , so we can always use the " SRTM 1Sec HGT (AutoDownload) "?

Raster data can be stored in only one coordinate reference system (CRS). Some are based on meters (e.g. UTM) while other’s units are degrees of latitude and longitude. You can set the output CRS of some operators or use the reprojection tool to change the spatial reference of a dataset.

About the firewalls: Fail in test connection - #2 by ABraun

1 Like

Thank you for your good replies .
I downloaded the External DEM but here I encountered this error .
Is it true that I have chosen External DEM? And why does “DEM No Data Value” have a zero value ?

error is “java.lang.NullPointerException”

Hi ABraun
isn 't that right ?

and The previous error was because I selected the data in size 30 and now I selected the size 5 , but this data does not cover my entire area . You don 't have any idea about that ؟

If you want to use an external DEM you have to provide it as geotiff in WGS84. You inserted a zip file, therefore SNAP doesn’t like it. If you want to go sure about its NoData value, you have to open it in a GIS and check, but 0 is mostly a good decision.

Your coregistration looks odd. If both inputs are reduced to the same bursts they should be contained in the coregistered stack as well. In your case, half of the product seems missing, but it’s hard to guess what went wrong.

1 Like