Coregistration: Sentinel-1 burst SLCs

Microwave Toolbox
1

I have several Sentinel-1 burst SLC images that I processed with:

apply orbit > calibration > deburst

When applying coregistration, I get the output file with all images, but only the master image has values; the remainder are empty.

I’m running the latest SNAP 13 on Windows 10.

2

Hi there!

I am still trying to understand how to handle the S1 bursts products, but I would believe that debursting is not needed on your chain.

Additionally, I am used to coregister SLCs without calibration, but let me ask you if your calibration provides you with complex output or just with sigma0/gamma0/beta0 values

Hope this helps.

Let me know

3

Hi, thanks for your reply!

You’re right, deburst is not needed because each file includes data from a single burst. But, I’m getting an error when trying to, e.g., coregister three burst files (where I applied first orbit + cal). The input files are the i and q from both polarisations.

”Error: [NodeId: CreateStack] For S1 TOPS SLC products, TOPS Coregistration should be used”.

But TOPS coregistration only allows me to coregister 2 SLCs, right?

If it helps, the burst SLCs were obtained from ASF and converted to SAFE format using a python package they developed.

Thanks

4

TOPS Coregistration (using only the Back-Geocoding operator) works also on individual bursts, please use it.

Also as previously done, if the TOPS-Split selected only 1 burst, the TOPSAR Coregistration using Back-Geocoding operator shall be used, but not the Enhanced Spectral Diversity, as this one only is applicable for multiple bursts product

The Copernicus Data Space Ecosystem also provides now single burst product and you can download it using the proper API or the phidown python package recently developed by Roberto Del Prete (GitHub - ESA-PhiLab/phidown: Manage, search, and download Earth Observation data with Φ-down from Copernicus missions with ease and efficiency.)

Best

5

Thanks!

I used the TOPS Coregistration with the back-geocoding operator and applied it to my multitemporal burst data. It works fine and generates the Is and Qs for the two polarisations and for each date.

However, my end product is to obtain a terrain-corrected covariance matrix (C2). If I try to calculate C2, I get replicated images of the master image.

One related note. I had a working processing chain with SNAP 12, which is no longer working. I processed all my burst data up to C2 terrain-corrected, and then I used the normal coregistration operator. In SNAP 13, this creates a multitemporal stack of only the C12 (real and imaginary) component of C2.

Thanks again for your support

Joao

6

Just wanted to point out this issue again (if it is one):

I processed all my burst S-1 IW data up to terrain-corrected covariance matrix (C2), and then I used the normal coregistration operator. In SNAP 13, this creates a multitemporal stack of only the C12 (real and imaginary) component of C2. The same processing chain with SNAP 12 produced the (expected) multitemporal stack of all elements of C2.

7

@lveci could you take a look?

8

Update: this happens not only with burst data but also with scene data

I tested it with the following 3 scenes:
S1A_IW_SLC__1SDV_20190702T092344_20190702T092411_027936_032774_4F50

S1A_IW_SLC__1SDV_20190714T092344_20190714T092411_028111_032CC0_4BC2

S1A_IW_SLC__1SDV_20190726T092345_20190726T092412_028286_033210_9CDB

9

Update #2: If I process from SLC to terrain-corrected intensity, then the co-registration runs alright. So, the problem in my case happens when using coregistration with terrain-corrected C2 data.