Dual Plane Imaging

1
Dual Plane Imaging
by Anna Moore INAF Osservatorio Astrofisico di
Arcetri, Florence
and Ian Parry Institute of Astronomy, Cambridge
2
Assumption 1
We can buy lots of (up to) 2K BY 2K NOISELESS,
PHOTON COUNTING detectors, 100Hz frame rate
L3CCD from e2v a good start (Craig Mackay
Near Diffraction-Limited Visible Imaging on
10-30m Class Telescopes with EMCCDs)
3
Assumption 2
We have access to several hundred TERRAbytes of
harddisk space and a large amount of CPU (dont
know how much CPU but for now it large)
4
(No Transcript)
5
Concept
Spatial sampling1,10,100 mas etc.
6
(No Transcript)
7
Move the detected intensity in each of the pupil
images so that they accumulate in only 1 pupil
image- note the transformation
Start here
8
1. transformation calculated and applied to
every frame 2. Add lots of frames together 3.
Sum pupil intensities for final optimised pupil
image
Faint target- not possible to detect under seeing
limited conditions
9
Why DUAL plane imaging? you can reconstruct
the focal plane image from the individual pupil
images (at least to the level of the spatial
sampling of the individual lenslets)
therefore every measurement gives both a
recording of the pupil plane and focal plane for
every spatial sample of sky, for 1 atmospheric
phase screen
10
Must have enough resolution elements across the
pupil images (not sure how few are required but
at least gt1 )
how does diffraction by the lenslet affect the
structure of the pupil image?

• Pixel size 15um (this is unlikely to change
  significantly)
• Largest detector size 2048 by 2048 (as above)
• Seeing disk diameter1 arcsec (governed by
  natural seeing of the site)
• Choose a sky sampling per lenslet (20mas, 100mas
  etc)

Calculate required focal length of telescope,
lenslet diameter
5. Input telescope diameter
Calculate input F/ratio to lenslet, focal length
of lenslet and hence airy disk diameter
Calculate pupil size/airy disk size of 1 lenslet
11
Telescope 2m diameter, 500nm
12
Telescope 8m diameter, 500nm
13
Telescope 30m diameter, 500nm
14
Telescope 100m diameter, 500nm
This is uncredible sparse sampling for a 100m
aperture- but with a lot of time and CPU how
well can one do? Can you reconstruct the
atmosphere?
15
And lastly, hard disk space
depends upon the detector format, but hundreds
of Terrabytes for 1 night
16
Conclusion
With noiseless detectors (and lots of them) you
can undo the effects of the atmosphere in
software rather than hardware

• Fully understand what were doing to the PSF
• Can choose between different algorithms to suit
  the science program
• Hardware required is extremely simple

17
THANKS FOR COMING