C2RCC: Retrieving Rrs from Sentinel 2?

Optical Toolbox
#1

Hello!

I have been using Sen2Cor and now recently C2RCC, which a colleague of mine finds matches very well with our field spectral data. However, I cannot reproduce his work, as I am unsure of the selections to make in the options window for C2RCC processing.

  1. I first resample all bands to 30m resolution.

  2. In C2RCC, I am choosing the product, and saving as BEAM_DIMAP, as well as checking these boxes:
    -output AC reflectances as rrs instead of rhow
    -output atmospherically corrected angular dependent reflectances
    -output uncertainties

All other items I leave standard.

  1. These are my results for Sentinel 2, displaying rrs4, rrs3, rrs2:

This is obviously (I think) wrong. Any help would be appreciated!

#2

I can’t see anything wrong. Actually I can’t see anything :wink:
The image is a bit too small. Could you show in addition an RGB image. Ah! Or is this an RGB?
Do you had a look ad the flags. Maybe they indicate that the result is not good.
What’s the value range of the rrs. Can you show a histogram?

#3

RGB_Sen2.png696×616 39.6 KB

#4

This (here I’ve enlarged it) actually is the RGB image of a lake(seen in white and yellow) surrounded by forest (seen in black with pink spots). The histograms for this “corrected” image are below in order of bands:

hist123.png644×716 128 KB
hist4567.png656×988 156 KB

I wonder if the sun glint is too severe in this image to provide decent results after corrections?
Glint in original image:

glint.png772×576 1.06 MB

There are also at least 20 flags, so my guess is that this correction failed through a bad image (glint?) or bad choices on my part.

#5

Yes, it can easily be that the glint prevents good results. When looking at the flags you should look especially at the Rtosa_OOS and Rtosa_OOR. Also if the Iop_OOR flag is enabled it is very likely that the AC is also not good.