I'm wondering what the best method is for correcting Sentinel-2 images downloaded in JP2 format (from Amazon Web Services - AWS) to BOA reflectance? As the files downloaded from the sentinel scihub in SAFE format are huge (usually about 6GB) and really cumbersome to process in the SNAP toolbox, I've been downloading the bands I need from Amazon Web Services in JP2 format, then converting them to layer-stacked GeoTIFF images in QGIS. But I need to correct them to BOA reflectance and don't know how best to do this.

Is it possible to use sen2cor in the SNAP toolbox with JP2/GeoTIFF images downloaded and processed from AWS?

  • I think SEN2COR expects you to open a metadata file rather than the images directly to get the required parameters to do the processing. I have had some luck using the USGS EarthExplorer which is less cumbersome than the scihub. Commented Nov 4, 2016 at 17:03
  • The safe format is specific and SNAP cannot process them if the metadata is not available. It is very inflexible in this way. I am not sure if the correct version of SNAP is available. ESA had an unfortunate snafu in updating their archive before the processing software. Commented Nov 4, 2016 at 17:36
  • I recommend the Semi-Automatic Classification plugin in QGIS. you can use it to search, download and correct your sentinel data
    – GeoMonkey
    Commented Nov 4, 2016 at 19:12

4 Answers 4


Some BOA products now may be downloaded directly from a S3 bucket of L2A data:


s3://sentinel-s2-l2a in EU Central region for L2A data over Europe, dating from 28th of March onward (the data is in a Requester Pays bucket which means that you can access it freely within the EU Central region, but you will incur charges if you download it elsewhere)


Some L2A images (BOA reflectance, with cloud mask) can now be downloaded from THEIA. They are produced with MAJA (not SEN2COR) and cover western Europe and some spots elsewhere.


Tools allow to download AWS images in SAFE format (https://github.com/kraftek/awsdownload) and then process them to surface reflectance (SR) with Sen2Cor in SNAP. There are few other methods such as ARCSI, Py6s, MACCS, MAJA, Sen2Agri, etc., but I haven't tried them. Sentinel Hub provides analytical SR data from Sen2Cor for a subscription.


You can use the sentinelhub python API to reconstruct the SAFE format from the JP2 files published to the S3 bucket. The library is published by the same group that maintain the Sentinel 2 S3 bucket.

I used this code to download files from S3 in a SAFE formatted directory

from sentinelhub.data_request import download_safe_format
download_safe_format(tile = (_tile, _date), folder='data', entire_product=True)

The API looks pretty rich so it may be worth investigating how to reformat files you already have to SAFE format.

I am able to successfully run sen2cor version 2.8.0 on the retrieved data.

Your Answer

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

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