Verify that adding higher-order Zernike from defocused images improves the focused image fit

This ticket is to verify that adding the Higher-order Zernike parameters estimated from defocused images actually improves the quality of the focused images, i.e., that the higher-order estimates are worth doing. 

One issue - how do we actually implement these high-order Zernike wavefront components into our analysis? At the moment for low-order components I simply conduct a linear fit of the Zernike amplitude as a function of defocus. It is not obvious that this will be possible for very high order Zernikes (due to possible degeneracies and low amplitudes). I will probably need to add a ticket dealing with this problem.