CR2CC algorithm

Optical Toolbox
#1

Hi all,

I am appying the C2RCC algorithm to retrieve the Kd value (resulting band Kd489 ). However, I need to automatize it and I have used GPT, but it takes a lot of time (around 1 hour only a “small” piece of the image). To decrease that time I wanted to save the pre-processed images (resampled, reprojected, clipped and with only the needed bands before. Can I store the pre-processed image in GeoTIFF or another format and then apply the C2RCC algorithm? Can I remove some bands before applying the algorithm or I need all bands to apply it?

Thanks in advance

#2

You can resample the data before, but you don’t have to reproject it.
GeoTiff should work, but probably it is better to use NetCDF4-BEAM or BEAM-DIMAP.
As input, are the radiance bands and the angle (sun and view) bands necessary.

If not already done, you can disable the computation of the uncertainties. This takes a lot of time.

#3

Please take a look at the following posts, they already discussed most of your demands,

The processing time

Source of the post

S2 Resampling Processor

Source of the post

C2RCC from the command line

Source of the post

For more information please take a look at,

EVOLUTION OF THE C2RCC NEURAL NETWORK FOR SENTINEL 2 AND 3 FOR THE RETRIEVAL OF OCEAN COLOUR PRODUCTS IN NORMAL AND EXTREME OPTICALLY COMPLEX WATERS
Source: https://www.brockmann-consult.de/wp-content/uploads/2017/11/sco1_12brockmann.pdf

1 Like