Fix join_segmentations dtype error for np.uint

[aeisenbarth]

Description

When the input arrays are of type uint64, adding a builtin type like 1 produces a float64. This change casts the result back to the input dtype.

Fixes #7291

Checklist

• A descriptive but concise pull request title
• Docstrings for all functions
• Unit tests
• A gallery example in ./doc/examples for new features
• Contribution guide is followed

Release note

Summarize the introduced changes in the code block below in one or a few sentences. The
summary will be included in the next release notes automatically:

...

[aeisenbarth]

I am wondering whether it's better to cast 1 to the matching dtype before doing the addition, rather than correcting the result afterward. If I knew the dtype, I could do np.uint64(1). But I don't know a way to cast a scalar to a variable type, like "1.astype(s2.dtype)".

Maybe double casting: np.uint(1).astype(s2.dtype)

[lagru]

Thanks @aeisenbarth! This already looks like an improvement.