I have to work with sentinel-2 data over 2016, my stydy area is located in France and it is covered by 4 tiles.
I have downloaded S2 L2A from https://theia.cnes.fr/atdistrib/rocket/#/home (which is very helpful).
Now i am wondering how to process this data, which steps i have to start with : stretching, mosaiking, reprojection , gap filling…
Did anyone has an idea and can advice me?
Thank you very much for your answer!
you forgot to mention what you’d like to do with the data.
Difficult to help you with such a broad question.
Thank you Olivier for your answer, I would like to use these data for mapping forest decline by computing vegetation index. My question was about the required preprocessing steps.
I want also to ask how to exploit the colud mask which is not binary.
Thank you in advance.
THEIA’s L2A are preprocessed, orthorectified and corrected from atmospheric effects. You may directly compute NDVI and monitor its variation with time. Just take care of the cloud mask and remove the cloudy pixels in your time series.
Thank you Olivier,What about the values of the cloud mask ? They are coded in 8 bits and we have the habit of a binary cloud mask. The value of 0 is clear it corresponds to no cloudy pixels but I’m wondering about the other values (1-255).
In the help page of theia server (https://theia.cnes.fr/atdistrib/rocket/#/help) you’ll find a link to the product documentation.
Our cloud and shadows mask also tells which threshold raised a flag, you have 8 flags in the mask.
But if you just want to separate valid and unvalid pixels, just test is the value is zero or not.