Is your feature request related to a problem? Please describe.
Especially older Sentinel-2 scenes can be mis-registered by several meters, causing a shift between scenes (Example). When conducting change analysis or time-series analysis at fine spatial scales, this will lead to misestimations.
Describe the solution you'd like
It would be very useful to support coregistration of a cube to minimize the shifts between different scenes. Either by aligning all scenes within the cube or using a base/reference image.
Describe alternatives you've considered
Other co-registration implementations exists, for example in the FORCE or eo-learn. However, when using xcube to request data from Sentinel Hub and to process the data further, I can't find a solution to make use of their solutions.
Hi @eikeschuett, this is a really good idea, but I fear currently out of scope for xcube. You can, however, plug in your own post-processors before generating data cubes from individual time slices.
Is your feature request related to a problem? Please describe. Especially older Sentinel-2 scenes can be mis-registered by several meters, causing a shift between scenes (Example). When conducting change analysis or time-series analysis at fine spatial scales, this will lead to misestimations.
Describe the solution you'd like It would be very useful to support coregistration of a cube to minimize the shifts between different scenes. Either by aligning all scenes within the cube or using a base/reference image.
Describe alternatives you've considered Other co-registration implementations exists, for example in the FORCE or eo-learn. However, when using xcube to request data from Sentinel Hub and to process the data further, I can't find a solution to make use of their solutions.