Copernicus
Industrial Grade Linguist
Like most questions in philosophy, there is a very simple answer. The problem is that philosophers disagree on the meaning of the question. If first they would all agree on the definition of the words physical and object then the answer becomes obvious.
Getting people to agree on meaning, even the meaning of what we think of as precise concepts, is illusory. We all have similar, but different, experiences of the world, so an expression like "physical object" is going to have a range of meaning that varies across contexts of usage. The best we can do is agree on broad slices of similar experiences when we debate the meaning of words. So we get a range of different reactions to questions like the one in the OP.
Unter has an interesting approach because of his focus on a perceptual description. We can contrast his approach with Bilby's, which focuses on a non-perceptual physical description. Both are reasonable ways to look at the problem, but from different ends of the range of components that go into our understanding of what a rainbow is. We have a physical model for describing the perception--rods and cones in the structure of the eye, photons that trigger neural activity, the subjective experience of color (or shades of gray).
The essential fact about a rainbow is that it cannot exist without an act of perception. There has to be a perceiver with the right sensory equipment. There has to be an external event that causes the perception (or, at a minimum, a hallucination--perception triggered by something other than incoming sensory input). Two different people standing next to each other can be said to see the same rainbow, but two people standing a kilometer apart will be seeing different rainbows (or none), depending of the refraction of light through the atmosphere. A camera can record the event and be used to reproduce an image of it later. So the phenomenon is definitely real, but it is also an illusion of circumstance.
The philosophical issue here is whether any physical object is fundamentally different from a rainbow. Remember that tables and chickens appear to exist even when nobody is observing them, but it would seem to be impossible for a rainbow to exist independently of an observer. So one is tempted to say that a rainbow is not a physical object in the same way, but it is still a physical object. So maybe BWE's "yes and no" response begins to make more sense.