NavList:

   

Reply

A few notes on possible camera issues when using them for CN:

- Rolling shutter might be a concern for CMOS sensors.  A CCD-based
DSLR doesn't have this problem (global shutter).  One way to get
around it is to shoot a pair of objects always at the same height in
the frame (y coordinate) so that the rolling shutter hits both objects
approximately simultaneously.  For collimation reasons one wants to do
this anyway.

- Shoot in raw mode.  I was pleasantly surprised that CHDK, the
third-party Canon firmware, supports the SX160.  The raw pictures are
a big improvement:  background noise is now uncorrelated, and the
images are 12-bit linear rather than 8-bit JPEG, so maybe a shade is
no longer needed for the Moon.  I'm also using just the green channel,
since the lens is showing some lateral chromatic aberration near the
edges.

- Image stabilization considered harmful?  After seeing spread of
nearly half an arcminute over a long series of shots that should be
much better, I'm starting to suspect it.  The next run will have it
off.  I believe this camera moves a lens element, not the sensor, and
perhaps this alters the distortion profile or has some other bad
effect.

- Use the bright sources.  I'm leaning back toward short exposures
(1/60 sec) and some a priori knowledge of which bright sources are
within the frame.  A pity.

- Use star-star distances as local calibrators.  For example, tonight
the Pollux-Moon, Betelgeuse-Rigel angles are nearly equal;
photographing them one after the other allows the sextant angle to
remain unchanged, cancelling all the sextant errors.  It's best to set
the sextant to the midpoint of the two angles; that way any
lens-distortion-model errors would also tend to cancel (I think).
Orthogonally to the above, one can also take Capella-Moon,
Capella-Aldebaran.  With these four photos, one should be able to pin
down the Moon in both coordinates.

Not there yet, though.  I should put everything on a tripod and use a
DSLR; that way there would be guaranteed success, and then work back
toward the handheld situation.

Cheers,
Peter

   

Reply