geo-deep-learning
geo-deep-learning copied to clipboard
Add quality control when creating hdf5 samples
Once sat and label data is written to hdf5, it becomes more opaque and difficult to verify. Minor misalignements and other defects, minor or major, between gpkg and source rasters have been noticed in the past.
How could quality control happen when preparing samples? Anyway to quickly view hdf5 samples once they are written, at least partly (ex.: on a random basis)? For example, a few samples (sat and label) could be rewritten with matplotlib.
This issue will expire the day GDL uses STAC items.
Once data is written to hdf5, it's difficult to verify their content. Could there be a quick way to do quality control (visualize with matplotlib, etc.)? It could be on a random basis (ie only part of the data could be visualized)
This issue expires the day we use STAC items of course.