I've left some inline comments. Probably establishing kernel normalisation via the test is the most important one.
PS: one thing I don't like about Astropy kernels is that they sometimes emit this warning:
WARNING:gammapy.detect.lima:Using weighted kernels can lead to biased results.
WARNING: The kernel normalization factor is exceptionally large, > 100. [astropy.convolution.core]
WARNING:astropy:The kernel normalization factor is exceptionally large, > 100.
This is currently present in our docs build, and should be silenced. Same here for SkyImage.smooth, we never should emit that warning just because someone has a large kernel. So either not use Astropy kernel smooth, or catch and ignore the warning.
I would put the width= lines below, instead of a separate if/else block here, even if it means duplicating one simple line for box and disk. But that's a matter of taste of course, if you prefer to leave it here, OK!
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.