I suspect the illusion comes from assuming the picure is realistic, in the sense that there is a place where one could stand and perceive the scene as depicted.

In my experience, which is limited to phone cameras and the like, they seem to produce realistic images at short ranges, but a landscape picture looks much smaller and more distant than how the scene seems to me as I take it, unless I zoom in. That has me wondering if there is some geometry that would give realistic-looking pictures at all distances? Perhaps if the camera is sized such that the lens-to-sensor distance is about the same as the eye-to-picture distance when viewing the images?

Another issue that I am wondering about comes from the fact that telescopes (and, I am assuming, telescopic lenses for photography) have an objective lens to produce an image that is much closer, and then an eyepiece (or whatever the equivalent is called in photography) to view that image. Does this have different perspective and foreshortening effects than a simple lens which merely casts an image directly on the sensor plane? To put it another way, for a given telescopic lens, is there, at least conceptually, an equivalent simple lens which would produce an identical image of any scene?