I’m processing Sentinel-3 data using batch processing from the command line (windows).
All is working out fine, the only problem is that in the output file folder of the single data (.SEN3.data) I can’t find the conc_chl data (it should be a .img file…). But when I open these output data with SNAP, they are right there, exactly in the ‘conc’ folder!
I can’t find a different way of explaining it, sorry!
This is the XML graph that I’m using.
graph_chl.xml (4.0 KB)
The two conc variables chl and tsm are ‘virtual’ data. They are computed on the fly within SNAP.
Check the properties of the bands and you see the expression.
More on the conversion of IOPs to concentration you can find at the C2RCC community page:
Neural Nets – C2RCC
If you want to have the stored into a file, you can save the product or a subset of it to a e.g. NetCDF4-CF file. Or (I’m not sure if this works) you can convert the virtual by right-click and then select ‘Convert Band’. Then save the product.
I tried to convert the virtual band and then save, and it worked.
Last question: it is possible to include this procedure in the graph? So the conc variables will be automatically converted and saved?
It is possible but not as easy as in the GUI.
First, you need to use the BandMaths operator to create a copy of the band. Simply us the name as expression.
Afterwards you need to merge (Merge operator) the output of the BandMaths with the original C2RCC. You need to exclude the virtual bands from the c2rcc and include the output from the BandMaths.
Attached is a sample graph which shows the BandMath and Merge operation in general.
math_and_merge.xml (1.1 KB)
Alternatively, you can directly use NetCDF4-CF as output format in your processing graph.
If you are only interested in the conc_chl you can also add a subset operation.
I made it! I adjusted the graph using the math and merge method. And it works!
Thank you so much for the hints!
Here’s the graph:
graph_chl_reproj.xml (5.1 KB)