search

dragonfly-science / csm-test

0 stars 0 forks source link

Introduction - what is meant by "cleaning" #2

Open katrin-berkenbusch opened 3 years ago

katrin-berkenbusch commented 3 years ago

@Fab-Doe can you please be more specific?

katrin-berkenbusch commented 3 years ago

"data processing"?

xycarto commented 3 years ago

'Cleaning' int his case would address re-alignment of the pixels to a specific extent, plus the introduction of 'Nan' values into nodata regions after reprojection. Specifically, we are cleaning up any errors that may have been introduced during the re-projection process. Cleaning can also refer to checking the metadata of the input Sentinel to ensure the information is in there. The second part is not as likley, since Sentinel is almost always ensured to be properly processed before it is avaikable to be public.

There is one more 'cleaning" that is performed as well. This is checking for cloud cover extent and 'cleaning' out unusable data before it reaches our process.

Happy to try and rework this to capture these processes?

katrin-berkenbusch commented 3 years ago

Yes, that sounds good as your explanation means that the term data processing is too generic perhaps.

xycarto commented 3 years ago

Updated to:

Nevertheless, implementation of the Sentinel-2 datasets pose a significant challenge for data ingestion, reprojection, pixel alignment after reprojection, and filtering cloud dense images. All of these are under investigation.
katrin-berkenbusch commented 3 years ago

thanks, I've edited it (and fixed the conflict). What is meant by "All of these are under investigation"? Is it part of the current project?