Hi,
I have this image (S2A_MSIL1C_20220515T081611_N0400_R121_T33HYD_20220515T102520.zip)
I am trying to run C2RCC. My workflow is: Resample->Subset->Idepix->C2RCC
The C2RCC results are very weird (attached image rhow_B4)
I am wondering why this black border appears after C2RCC.
This image is the Band 4 just before I run C2RCC
Any ideas?Thanks in advance
Lazaros Spaias
You can try to do the process with the graph builder in one shot and see if you have the same problems. I noticed those missing holes before. Usually it is recommended to save the products before using them as input of the next function. I am not sure if that is what happens, but it could be.
Hi,
In the example I run the workflow step by step manually just to recreate it . I get the same result when I run it headless with gpt. 
Do you mean that it only happens in the last step, when applying C2RCC?
Yes, correct. I am explaining it in my post as well. I have an image screenshot before C2RCC and one after. The problem with the black border appears at the C2RCC step
It looks like this is related to the resampling. I guess you have used the general resampling tool.
This does not cover the specialities of S2 data.
Please try the one which is in the menu at Optical / Geometric / S2 Resampling Processor.
As you can see in the attached screenshot the view angles are not correctly resampled by the generic resampler (the left side). This causes C2RCC to fail in these areas.
This worked! thanks a lot
Cheers,
Lazaros
This could be possible but is not a good design. A dependency from the generic code to the specific is generally not good. Even if it would be a weak dependency in this case. But the generic tool would need to have some knowledge about S2. For the short term it might work.
Actually, it would be better to integrate and generalise the S2 Resampler in the Generic Resampler.
Then there would remain only one resampling processor.
This has been discussed years ago but wasn’t implemented.
Alternative draft - SNAP - Confluence (atlassian.net)
