Use smarter default contrast settings for source thumbnails

[timstaley]

CF https://redmine.astron.nl/lofar_issuetracker/issues/4616

(NB this is a separate, and perhaps much easier issue to #52 since we're just choosing the static defaults)

Jess suggests using a multiple of the local RMS, but for that to work we'd have to record the local RMS... so that's a non-starter. Is the Gaussian fit error approximately the same thing? I'm not sure, but it could be a good value to start playing around with.