What about the type of imagination which medicine calls hallucinations? Is it a great gift or a curse for those who sense extremely? Is imagination a mild form of hallucinating?
There are certainly some anecdotal stories that hallucinations (as part of psychosis) occured in highly creative people e.g. van Gogh or in the families of some creative people, such as the story of James Joyce's daugher. But that's anecdotal evidence. I tried to find whether some (psychometric) measure of imagination correlates with hallucinations, but not much luck with that insofar. [And that's not because such measures of imagination don't exist at all; some have been proposed. I'll update this part if I find something relating them to hallucinatons.]
On a slightly more positive note, the propensity for hallucinations was also associated with a better pattern matching ability in some visual tests with highly confusing images, but this was only when exposed to the full image before, i.e. a form of bias toward prior knowledge.
Perhaps the above aren't so surprising given that a lot of normal people (about half it seems) hallucinate in sensory deprivation experiments. A somewhat similar situation was observed in people who experienced hallucinations as a side-effect of higher doses of tricyclic antidepressants, namely that their hallucinations occured when background noise was low. Charles Bonnet syndrome is anohter case of reduced stimuli; Oliver Sacks has a talk at TED about it. So if anything, hallucinations are associated with not enough perception.
Note however that there's a rich tradition in philosophy of discussing imagination and hallucinations together. E.g. there's a MIT Press book titled Hallucination: Philosophy and Psychology with quite a few hits inside for "imagination" So it's quite possible I've missed some relevant issue. Do note as a caveat that one of the papers (by K. Farkas) in that book mentioned early on that
The notion of hallucination most commonly discussed in philosophy is somewhat different from the notion used in psychology or psychiatry.
Likewise there's a Humanities paper that discusses them together. (Despite the dubious venue in which it was published, it is cited by papers in more reputable journals, so perhaps worth a read--but it's 50 pages, so don't expect me to do it.)