C2RCC for processing S2 data over complex turbid water

Hi Im new to SNAP. I am trying to use C2RCC with C2Rcc nets for processing of S2 images using SNAP 6. Can I use this model for correction over complex turbid waters or should I use any other model Pls suggest.
How can I use C2CRR in batch mode.

The documentation for the C2RCC is still in progress. So there is no detailed documentation as you have asked in the other post.
If you have very turbid water you should the so called ‘C2X-Nets’ set of neural nets.
For using the C2RCC from the command line, you can have a look here.
Also this might be helpful for you: Bulk Processing with GPT

thank you Marco

Hi
when i run C2CRR on landsat 8 OLI data it reduce the image coverage (I mean full image is not being processed)
in picture below is processed data using 6S v.1 and above one is C2CRR .
why its so? can u help?

This is caused by the bad flagging in L8 data.
By default C2RCC processes pixel where water confidence is either medium or high.
At the borders, it is sometimes not even low. Also, other flags (like vegetation confidence) are not really set appropriate.
You change the valid pixel expression to something more suitable.

Dear Marpet
The C2CRR is taking so much time such a 2-3 hours for processing area half a tile of sentinle-2 coverage.
Is is the issue of my system (I am using i7) or it usually take such time ?
As I have to process large amount of data so what is the possible solution ?

The processing time depends on which resampling you are using. The general one or the one specific for S2. The general one is faster, the specific one a bit more accurate. Also the amount of memory your system influences the processing speed. And finally, the selected processing parameters can stretch the processing. For example, disabling the output of uncertainties will speed up the processing significantly.
I have processed the following scene in less than an hour:
image
No uncertainties, general resampling to 10m and my system has 16GB of RAM, 11GB assigned to SNAP.

1 Like