Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Pixel Value Discrepancies in Sentinel-2 L2A COGs Compared to Other Distributions #56

Open
wakame1367 opened this issue Dec 11, 2024 · 1 comment

Comments

@wakame1367
Copy link

Description:

We've observed significant differences in pixel values between Element84's Sentinel-2 L2A Cloud-Optimized GeoTIFFs and other distributions of the same scenes, specifically:

  1. Microsoft Planetary Computer's Sentinel-2 L2A
  2. Original JP2K files from AWS Sentinel-2 registry (s3://sentinel-s2-l2a)

Observation Details:

  • The cloud masks and general scene geometries match perfectly at pixel level, confirming the source data is identical
  • However, raster histogram analysis in QGIS shows different pixel value distributions
  • Notably, the original JP2K files in AWS (eu-central-1) match with Microsoft Planetary Computer's distribution
  • The discrepancy appears to be specific to the COG versions

Verification Steps:

  • Compared identical scenes across all three sources
  • Used QGIS raster histogram analysis to verify pixel value distributions
  • Confirmed that cloud geometries match at pixel level
  • Verified that original JP2K assets show consistent values with Microsoft PC

Questions:

  1. Are there any known processing steps during COG conversion that might explain these differences?
  2. Could there be any parameters in the COG conversion pipeline that might be affecting the pixel values?
  3. Is this a known issue or expected behavior?

We want to ensure we're not missing anything in our analysis before proceeding with using these datasets.

Technical Details:

@piyushrpt
Copy link

Which collection are you using for the comparison? What type of differences are you seeing - are they off by a constant?

You should be looking into the sentinel-2-c1-l2a collection if you want to check consistency against original JP2 files.

It is well documented that sentinel-2-l2a adjusts the imagery to account for the offset introduced by ESA's new processing baseline.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants