Radiometric & Geometric Correction Workflow

[ABraun]

as water bodies are always flat and have low backscatter (excluding waves, ice cover or oil/sand particles) I doubt that terrain flattening has a larger effect.

Regarding the virtual band: Right click > convert band and then menu > file > save.

[clardeux]

Just a suggestion but it should be good after orbit to apply thermal noise removal in order to decrease radiometric variation between subswath.

Be carefull because it’s usefull for data acquired after the end of november.

Best

[jvcc]

polarimetry only works on the i and q parts of the image

Can anyone confirm whether this is the case and a flattened image can’t be processed for polarimetry?

[lveci]

Currently a flattened image is only producing gamma0 intensities.
We need to see how we can change it to allow you to still create a
polarimetric matrix.

[jvcc]

Just a suggestion but it should be good after orbit to apply thermal noise removal in order to decrease radiometric variation between subswath.

Be carefull because it’s usefull for data acquired after the end of november.

Do you mean image improvement from thermal noise removal is most important during fall/winter months? Excuse my ignorance, I haven’t had the chance yet to look into thermal noise removal at all.

[clardeux]

Sorry, if I said after novembre it’s just because calibration of sentinel 1 was improved at the end of 2015 november.

Consequently, after this time the thermal noise metadata is now relevant.

I hope now my answer is more clear…
Best

[ABraun]

thanks for the clarification. Do you have any source where I can read about this improvement?

[clardeux]

Yes, it’s here

https://sentinel.esa.int/web/sentinel/news/-/article/improved-radiometric-calibration-of-sentinel-1-tops-products

[Phil]

Hey together,

i want to preprocess Sentinel-1 SLC images for a subsequent agricultural classification.
I am an absoulute beginner in working with SAR-Data, so i would like to apologize for any bad mistake.
My suggestion for the preprocessing steps and the order:

1. Radiometric calibration
2. Single Product Speckle Filter
3. DEM Assisted Coregistration with XCorr
4. Range-Doppler Terrain Correction
5. Spatial Subset to my Study Area

does this make sense?
I would be very grateful if you would add some improvements!

Thanks in advance,
best regards,
Phil

[ABraun]

are you sure you are using SLC data? In this case you would need to perform TOPSAR Deburst after Calibration in order to remove the black stripes.

Coregistration is only needed when you work with multiple datasets at once.

You could also make a subset earlier in order to save computing time .

[Nzx]

I try to obtain the backscatter coefficient of Sentinel-1A using SNAP. Is the radiometric calibration the first step?

[ABraun]

depends on the level of your product. Have you downloaded SLC or GRD data?

[Nzx]

SLC, IW mode, VV and VH polarization

[Nzx]

If I perform TOPSAR Deburst, Filtering and radiometric calibration, the results of VH polarization is Ok, but the values of VV polarization are zeros.

[ABraun]

for SLC data I would suggest

1. TOPS Split
2. Apply Orbit file
3. Thermal Noise Removal
4. Calibration to Beta0*
5. TOPSAR Deburst
6. Radiometric terrain flattening
7. (Speckle filtering)
8. Range Doppler Terrain Correction

If you don’t need the phase information you can also download it as a GRD product and only apply the following:

1. Apply Orbit file
2. Thermal noise removal
3. Calibration to Beta0
4. Radiometric terrain flattening
5. (Speckle filtering)
6. Range Doppler Terrain Correction

or (if you don’t have much topography):

1. Apply Orbit file
2. Thermal noise removal
3. Calibration to Sigma0
4. (Speckle filtering)
5. Range Doppler Terrain Correction

[Nzx]

Thank your very much for your suggestions. Does the Thermal noise removal and Apply Orbit file are the necessary steps?

[Nzx]

I performed the SLC data using SNAP
1 TOPSAR Deburst
2 Calibration to Sigma0
3 Speckle filtering
4 Range Doppler Terrain Correction
For the radiometric calibration results, the VV is right, but the VH is incorrect.

[ABraun]

Thermal noise is not present in every product. In case you don’t rely on highly accurate radiometric values you can also skip it.
The orbit file adds information about the geometric location of your product. If you want to combine it with other data it can improve the position accuracy.

Your suggested workflow is sufficient for most applications.

Why do you think calibrated VH is not correct?

[Nzx]

Thanks for your reply, I follow your suggestion to perform the Sentinel-1A data using SNAP. In the previous results, when the radiometric calibration was performed, the results of VV polarization are zeros, only the VH polarization is correct. Can you supply an email for me, thus I can transform the results to you.

[ABraun]

sorry, I can’t help you outside this forum.

But indeed, zeroes as an output are obviously wrong. Was the data correct/complete before radiometric correction?