Hi! I’ve noticed that C2RCC generates an output mask called VALID_PE (with a pre-defined valid pixel expression, set by default as B8 > 0 && B8 < 0.1 ). Is this mask derived from the raw image? Can I use it in the raw image to maintain only valid pixels?
Also, I know that Sen2Cor generates an output mask called quality_cloud_confidence… I want to join those 2 flags in a same image (VALID_PE and quality_cloud_confidence) in order to improve invalid pixel removal. Have anyone done this? Do you think it is a good idea?
C2RCC has also a Cloud_risk mask (quite conservative) that you could check, and compare with the Sen2Cor one. Anyway, you can transfer mask bands to other products using the mask manager, click on this icon:
Thank you so much! I’m going to test Cloud_risk mask then (since it’s a cloud mask, it is derived from the raw image. But…do you know whether those other C2RCC quality flags (Rtosa_OOS, .Rhow_OOS etc) are applicable as quality flags to other AC processors output (Sen2cor and iCor)? (in order to filter valid data) @abruescas
Welcome! Rtosa_OOS, Rhow_OOS are the Out Of Scope flags of the NN training dataset, that is, radiances that are not considered valid by the C2RCC NN AC part. I think you could transfer to Sen2Cor and iCor to have the same valid pixels in all three AC processors for comparison procedures. But after testing you could also see that some of the values are reasonable and decide not to use them. Sometimes they flag too much…I guess you will have to check and make decisions depending on the scenes/cases.
Thanks again! I’m going to use test those C2RCC quality flags on both Sen2Cor and iCOR for comparison procedures and check