Add Douglas-Peucker line simplification for morphologies

[eleftherioszisis]

Looks nice, I didn't go through the algo, nor check how it modifies morphologies, but i think you can merge, and maybe just add a note that this is 'experimental', so users don't try blindly and then come back complaining ;)

Experimental in what sense?

[arnaudon]

Like which values of epsilon are good, and no pointer to a published results/paper using it and showing in which context this is useful, and possible impact on the morphology tortuosity, etc...