Normalization of L1C Data

#3
by pmcelroy - opened

From the example given in the project overview, it appears data is normalized by a factor of 5000 (as opposed to 10000) before visualizing:

datapoint_rgb = np.moveaxis(datapoint[[3, 2, 1]], 0, -1) / 5_000

Where does this factor of 5000 come from?

Also, I'm assuming all samples are from pre-2022 (i.e. before any offset was introduced on Sentinel 2 products)?

Thanks in advance!

Image and Signal Processing • ISP org

Hi, it's just for visualization you must divide by 10_000 to obtain TOA values.
The data is retrieved from https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_HARMONIZED

csaybar changed discussion status to closed

Sign up or log in to comment