Changed s2dv::MeanDims by multiApply(fun = mean) for not losing the dimnames

Hi @nperez,

I have opened this merge request to fix a bug that arises because the name of the dimension is lost (when the output only has one dimension) when using the last version of s2dv::MeanDims().

I only changed MeanDims() by Apply(fun = mean) in two lines of Calibration(cal.method='rpc-based').

Please let me know if you think it makes sense to fix this bug in another way.

Thank you!

Carlos