It is simply not necessary to take the pitch of the camera into account, as long as the picture taken includes the relevant parts, ie the horizon (and those pictures show a reasonable horizon for sure.)
Let’s use Toms current favourite, the Bishop Jellybeans If you line up the jellybeans, and look at floor level along them, they will all line up, and superimpose. Or are in Transit. That is his statement and i guess he will need to stick to that.
Now without moving your head (camera) from the plane of the jellybeans, (the floor) and glance up at a point a few degrees above the far jellybean. Do they come out of alignment? Of course not. Now do the same, moving your eyes only, (camera lens) glance down slightly at as point a few degrees below the jellybeans. Are they still in alignment or transit? Of course they are.
All, you have done is tilt the camera lens (eye) on the axis of the plane, and altered the angle of the lens slightly, allowing you to focus slightly above or below the jellybeans. They are still in alignment or transit, they have not moved, and niether has what you see (apart from more of what is above or below the jellybeans)
There is no way in that experiment to get the water in the nearest tube to align with the water in the furthest tube AND align with the horizon, as the horizon is on a different plane.
You can always get the level of 2 elements to align, but never the 3.