You are assuming a model of perspective which stretches into infinity continuously without meeting, rather than the finite perspective model observed where the lines appear to meet.
I am asking why you are waving your hands frantically and asserting that there is an unobservable stretch of infinity where the perspective lines meet. You are sounding like someone who is trying to convince others of the existence of ghosts.
What we're assuming is that if we can see a 1m wave which is, say, 5 miles away on the horizon then I can also see the SODDING GREAT BUILDING which is a few meters beyond it. I'd like to see you or someone else show a diagram demonstrating how a ray of light from a 100m building can be blocked by a 1m wave unless the wave is close and your eye is below 1m.
How perspective works in the real world has been explained to you multiple times. The size you perceive something depends entirely on the angle the light rays from either end of it meet at your eye. The light from either track at A meets at a bigger angle than the light from either track at B, thus you perceive A as being bigger than B even though in reality they are the same size:
That's it. At what distance does that angle become 0, given that you can see this is a triangle? At infinity.
BUT, we don't have to wait till infinity, in real life we can't perceive things if that angle is below a certain size.
That is where the rails would actually appear to meet at a finite distance simply because we could no longer distinguish them.
Optical zoom would then restore them so we can perceive them as separate rails again because zoom increases that angle again such that we can perceive things more clearly.
But, even without optical zoom, if there was a big building beyond where we can no longer distinguish the two rails then of course we could see it, it's a lot bigger than the distance between the rails, the light from the top and bottom meets at our eye at a much bigger angle than that of the two rails so we can clearly see it.
It's interesting how objects being occluded by the horizon (which in the real world is the edge of the earth) are explained by merging perspective lines or things on the horizon blocking them - ergo, it's not because the earth is curved. But then long distance photography is used as proof that the earth is flat because if the earth is curved then you shouldn't be able to see them. Well...if your model of perspective is correct then you shouldn't be able to see them either. You can't have it both ways...