MPM binary_closing

[amoodie]

I was messing around with the MPM because it was very slow on a large domain.

DeltaMetrics/deltametrics/plan.py

Lines 1771 to 1775 in 999bbe9

	while (_changed != 0) and (_iter < 100):
	_iter += 1
	_newimg = morphology.binary_closing(img, footprint=disk)
	_changed = np.sum(_newimg.astype(float) - img.astype(float))
	_closed = _newimg

@elbeejay (sorry I think you wrote this bit) Am I correct in reading the lines here that the image input to binary dilation is not changed on each iteration? This seems like a bug?

[elbeejay]

Yes, looking with fresh eyes it does seem like this loop will never resolve short of hitting the iteration limit. But given that we iterate through the disk sizes in the public function, I'm struggling to remember why iterations here were ever needed wanted - seems like it'd be simpler to just call the binary closing function once and be done (in which case why use a private function at all).