I’m in Ireland for the week, at the invitation of the Institute of Physics Ireland, giving talks at a few universities here. My first stop was at the National University of Ireland, Galway. In addition to the pleasure of spending a bit of time in this very attractive country, it’s always interesting to get a chance to learn what people are doing in the departments one visits. The physics department at Galway is small, but it’s received a lot of investment recently; the Irish government has recently started spending some quite substantial sums on research, recognising the importance of technology to its currently booming economy.
One of the groups at Galway, run by Chris Dainty, does applied optics, and one of the projects I was shown was about using adaptive optics to correct the shortcomings of the human eye. Adaptive optics was originally developed for astronomy (and some defense applications as well) – the idea is to correct for a rapidly changing distortion of an image on the fly, using a mirror whose shape can be changed. Although the implementations of adaptive optics are very sophisticated and very expensive, we’re starting to see much cheaper implementations of the principle. For example, some DVD players now have an adaptive optics element to correct for DVDs that don’t quite meet specifications. One idea that has excited a number of people is the hope that one might be able to use adaptive optics to achieve better than perfect vision; after all, the eye, considered as an optical system is very far from perfect, and even after one has corrected the simple failings of focus and astigmatism with glasses there are many higher order aberrations due to the eye’s lens being very far from the perfect shape. The Galway group does indeed have a system that can correct these aberrations, but the lesson from this work isn’t entirely what might first expect.
What the work shows is that adaptive optics can indeed make a significant improvement to vision, but only in those conditions in which the pupil is dilated. As photographers know, distortions due to imperfections in a lens are most apparent at large apertures, and stopping down the aperture always has the effect of forgiving the lens’s shortcomings. In the case of the eye, in normal, daytime conditions the pupil is rather narrow, so it turns out that adaptive optics only helps if the pupil is dilated, as would happen under the influence of some drugs. Of course, at night, the pupil is open wide to let as much light as possible. So, does adaptive optics help you get supervision in dark conditions? Actually, it turns out that it doesn’t – in the dark, you form the image with the more sensitive rod cells, rather than the cones that work in brighter light. The rods are more widely spaced, so it turns out that effectively the sharpness of the image you see at night isn’t limited by the shortcomings of the lens, but by the effective pixel size of the detector. So, it seems that super-vision through adaptive optics is likely to be somewhat less useful than it first appeared.