ASTROPHOTOGRAPHY - TIPS Last updated: 11 July 2007

Subject:	Dither: a way to improve imaging resolution
Sent:	Tuesday, July 10, 2007 14:10:06
From:	Theodore Rafferty (raff650@verizon.net)
Dither
 
In the piece I wrote on Drizzle alignment (see:
http://www.weasner.com/etx/astrophotography/2006/drizzle.html), I said
that Drizzle works best if the stars are shifted between the frames that
are to be aligned and combined. That if a series of exposures are taken
with perfect guiding that Drizzle alignment will have no affect. Usually
there is some random guiding errors or field rotation (caused by the
polar alignment being off some and the guide star being used not being
close to the target being imaged) that causes small shifts in the star
positions on each frame, but these conditions can vary and are not
something you can easily control. A way to insure there is a shift is to
move the telescope in both RA and Dec slightly. Moving a telescope in
this way to aid Drizzle is called "Dither" (both Drizzle and Dither were
developed at the same time by NASA for taking and processing images from
the Hubble Space Telescope).

To illustrate how Dither can help Drizzle, I took two series of
exposures of M57 using a Meade LXD75 8-inch f/4 SN and a Meade DSI Pro 2
imager. For the first series, I took ten guided 30-second exposures. For
the second series, I took five guided 30-second exposures, then moved
the telescope slightly in RA and Dec (i.e. Dither), and then took five
more guided 30-second exposures.

The first image below is a combination of the first series' ten frames
with no attempt to align them to each other using the stars.



The image is an enlarged section of the full image to better show what
is happening. The guiding wasn't perfect and you can see a slight
elongation left to right (likely caused by the guiding not correcting
the motion cause by the periodic error in the RA quickly enough).

The next image is a combination of the ten frames from the first series
that have been Drizzled aligned (the original 748 by 577 frames where
enlarged to 1496 by 1154 and a two-star alignment was preformed) and
averaged together.



The slight drift in the frames from left to right allowed Drizzle to do
a pretty nice job of improving the resolution.

The next image is a combination of the ten frames from the second series
(where the telescope was Dithered in the middle of the series) with no
attempt to align them to each other using the stars.



Because the telescope was Dithered in the middle of the series, all the
stars appear as doubles.

The last image is a combination of the ten frames from the second series
that have been Drizzled aligned and averaged together.



The resolution is slightly better for the Dithered series compared to
the non-Dithered series. Plus Dithering the telescope insures a shift of
the field between frames that in turn provides Drizzle something it can
work with.

Using Dither and Drizzle are particularly helpful for those using focal
lengths and pixel sizes that result in undersampling, such is the case
with the equipment I used for these images. Undersampling is when the
focal length and pixel size combination produces star images that cover
too few pixels and they appear to be made up of blocks instead of
appearing round. This is most common for short focal length optics used
with imagers with large pixels. Oversampling, on the other hand, is when
the focal length and pixel size combination produce star images that
cover more pixels than are necessary. This is most common for long focal
length optics used with imagers with small pixels. By combining Dither
and Drizzle is a neat trick to improve resolution from less than ideal
equipment.
 
Ted Rafferty
Gaithersburg, Maryland