Future prospects

There are a number of modifications to the Lucky Exposures method as presented here which would make it applicable to a wider range of astronomical observations. In crowded fields, the image quality could be monitored using a number of reference stars across the field, allowing exposures to be selected on the basis of isoplanatic angle as well as overall image quality. By combining data from several different reference stars, the signal-to-noise ratio for Strehl ratio measurements could also be improved.

The Lucky Exposures method is not restricted to single-wavelength detectors - light from a science target could be directed into a spectrograph with an Integral Field Unit (IFU) while the light from a reference star was monitored on a conventional imaging detector in order to select moments of high image quality. Array detectors with spectroscopic sensitivity such as Superconducting Tunnel Junction (STJ) devices could also be used to provide spectral information. If the reference star is faint, a broader bandpass could be used for the reference star than for the observations of the science target.

The dependence of $r_{0}$ on observing wavelength described in Equation 2.9 implies that the Lucky Exposures method should work well on much larger telescopes if longer observing wavelengths are used. An $8$ $m$ telescope observing at K-band would have the same number of $r_{0}$ across its diameter as a $2.5$ $m$ telescope observing at $800$ $nm$ wavelength, and a similar probability of Lucky Exposures would be expected. Current low noise infra-red cameras can typically only be read at low frame rates, so further camera and detector development might be required to make such an instrument viable. Observations could also be performed at shorter wavelengths using smaller telescopes, although this would probably require faster camera readout rates and possibly an atmospheric dispersion corrector.

In order to improve the resolution attainable with the Lucky Exposures technique, non-circular apertures could also be exploited. If a large (diameter greater than $7r_{0}$) telescope were broken up into a series of slit apertures, the probability of obtaining good atmospheric conditions over one of these slits would be higher than for the telescope aperture as a whole. By repeating observations with a range of different slit position angles, high resolution data could be obtained in all orientations from an astronomical target.

Alternatively, a low-order adaptive optics system designed for long wavelength imaging might provide a substantial improvement to the probability of obtaining Lucky Exposures at short wavelengths on a large telescope, as it would eliminate the large scale structure in the atmospheric phase perturbations. This could allow high resolution imaging from large telescopes without the need for high-order adaptive optics correction (which usually requires a bright reference star).