I am trying to understand the data that is provided in EuroSAT : Land Use and Land Cover Classification with Sentinel-2. It is comprised of "Sentinel-2 satellite images covering 13 spectral bands and consisting out of 10 classes with in total 27,000 labeled and geo-referenced images". According to the paper the multispectral images have a resolution of 64x64 pixels each.

Table 1 - All 13 bands covered by Sentinel-2’s Multispectral
Imager (MSI)

What I don't understand about this is the spacial resolutions (Table 1 in the paper) of the different bands that range from 10 meters to 60 meters. Under the assumption that each band has 64x64 pixels, wouldn't that mean that the bands are covering different geographical areas?

If that were the case, how would you work with such kind of data for tasks like image classification or segmentation? Would you somehow crop the images or rescale them to fit the same spatial area?


1 Answer 1


See page 6 of the paper you link to:

Bands with a lower spatial resolution have been upsampled to 10 meters per pixel using cubic-spline interpolation

  • Thanks, I missed that @user2856
    – 94621
    Nov 1, 2021 at 12:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.