Estimate seeing conditions

[bgottula]

Implement an estimator for the seeing conditions. I think the standard way this is quantified is by the Fried seeing parameter r0, which can be estimated by taking a somewhat longer exposure of a star and estimating the half-magnitude diameter of the resulting disk.

[bgottula]

There may be some overlap with autofocus work in #248 since that must estimate the half-flux radius of a star. This is slightly different from full width half magnitude but is more robust under poor seeing, so it might actually be a good estimate for seeing conditions.