Today I forgot to clear the user filter in OSMAnd and I redid a lot of perfect and recent work done by bhiker.
Sometimes i wonder whether having some computer vision learning to detect ‘poor quality’ images could mean that ‘better’ images are promoted.
But I can also imagine how difficult it would be to get it consistent and correct.
The same applies to an idea that myself, and others, have had, which would have older sequences gradually demoted in favour of more recent captures. But just because it is taken more recently, does not guarantee and improvement.
It appears there are lots of imperfect options. But I still believe that more coverage - even of areas that have been covered, is good.