How do radio telescopes form raster images?

[Peter Webb]

In optical telescopes, the parabolic mirror forms an image on a CCD. For a
radio telescope operating in (say) the 1 cms band presumably there is some
matrix of receivers 1 cms apart to form a raster image, to save them having
to do each pixel separately to form a high res image, is that how they work?

[Bill Owen]

Not exactly. A radar image is a three-dimensional plot of received
intensity as a function of Doppler shift (on the x axis) and time delay
(on the y axis).

You transmit a brief burst of energy at some frequency. It first hits
whatever spot on the body is closest to you (the body's "nose" as it
were), and then it hits the rest of the body. So the echo from the nose
is received first, then the echo from the rest of the body. Meanwhile,
the echo has a Doppler shift impressed upon it from the radial velocity
of each piece of the surface. The side of the body which is rotating
toward you will produce a blue-shifted echo relative to the side which
is rotating away from you.

The result looks like a picture that's being illuminated from below, but
it's not the same as taking a picture. The "distance" in y, once you
convert it from microseconds to km, does give you the radius of the
body. The "distance" in x measures the rotation rate. It is possible
to see features in a radar image, and to track them as they move across
the front face of the body, and thus we can get the rotational period
independently of the spread in the Doppler.

It's also fairly easy to detect binary objects. The orbital velocity
will produce different Doppler shifts for the two components, and the
difference in range will separated their signals in time.

Steve Ostro was one of the leading experts here at JPL before his
untimely death. Lance Benner and others have taken up the task.

-- Bill Owen

[Steve Pope]

The Wiki article on Focal Plane Arrays addresses the radio telescope
(as opposed to radar) cases pretty well. The short answer is you
can have an array of detectors in the focal plane, or a phased array
to do the same thing.

In an active emitter system like a radar, particularly if it can
rotate and/or is orbiting (say for earth/planetary observation), you
have a lot more design freedom than a radio telescope, and you can form
more of a complete 2-D or even 3-D image.



Steve

[Martin Brown]

On 30/03/2012 22:08, Bill Owen wrote:
> Peter Webb wrote:
>> In optical telescopes, the parabolic mirror forms an image on a CCD.
>> For a radio telescope operating in (say) the 1 cms band presumably
>> there is some matrix of receivers 1 cms apart to form a raster image,
>> to save them having to do each pixel separately to form a high res
>> image, is that how they work?
>
> Not exactly. A radar image is a three-dimensional plot of received
> intensity as a function of Doppler shift (on the x axis) and time delay
> (on the y axis).

Not at all would be much closer.

Radio telescopes of the steerable big dish type like Jodrell Bank and
others like it use a technique called basket weaving which is in effect
measuring signal as a function along a raster grid in a particular way
that helps eliminate systematic errors. Typically there are a pair of
receivers and a difference signal recorded.

http://ned.ipac.caltech.edu/level5/March01/Andernach/Ander2.html

Most radio telescopes are entirely passive devices and almost all
serious ones at high resolution use Earth rotation aperture synthesis to
measure visibility fringes over a series of baselines between a set of
accurately surveyed dishes and from that deduce the sky brightness. The
observations are the Fourier transform of the sky brightness and the
image is obtained by computation after the observations.

NRAO has a reasonable basic introduction for the OP

http://www.nrao.edu/index.php/learn/radioastronomy/radiotelescopes

There are a lot of tricks used in practice to avoid seeing ground based
interference as phantoms in the sky.

--
Regards,
Martin Brown

[Steve Willner]

In article <JgDdr.25735$cd7....@newsfe06.iad>,

Martin Brown <|||newspam|||@nezumi.demon.co.uk> writes:
> Radio telescopes of the steerable big dish type like Jodrell Bank and
> others like it use a technique called basket weaving which is in effect
> measuring signal as a function along a raster grid in a particular way
> that helps eliminate systematic errors. Typically there are a pair of
> receivers and a difference signal recorded.
>
> http://ned.ipac.caltech.edu/level5/March01/Andernach/Ander2.html

A few radio telescopes are now equipped with focal plane arrays, as
another poster mentioned. I believe the one at Arecibo has seven
receivers, for example. That speeds up mapping speed but doesn't
create much of an image in a single pointing.

> Most radio telescopes are entirely passive devices and almost all
> serious ones at high resolution use Earth rotation aperture synthesis to
> measure visibility fringes over a series of baselines between a set of
> accurately surveyed dishes and from that deduce the sky brightness. The
> observations are the Fourier transform of the sky brightness and the
> image is obtained by computation after the observations.
>
> NRAO has a reasonable basic introduction for the OP
>
> http://www.nrao.edu/index.php/learn/radioastronomy/radiotelescopes

I did a quick Google search on "aperture synthesis interferometry"
and didn't come up with anything I liked very much. Anyone
interested should try it, though. There are lots of pages on the
subject, and maybe one of them will appeal, or maybe a combination
will be effective. It's a hard subject to explain, though, without
making lots of diagrams and pointing at them. The basic idea is to
use multiple antennas observing at the same time and linked together,
but there is (normally) only one receiver per antenna. The Very
Large Array (VLA) with 27 antennas is probably the best known
telescope that works this way. An extension is Very Long Baseline
Interferometry, where the telescope linking is via extremely accurate
clocks at each telescope site rather than physical connection.

The classic "student text" is _Synthesis Imaging in Radio Astronomy_.
I have the 1988 edition, volume 6 in the ASP Conference Series, but
I'm pretty sure there are later versions. The basics won't have
changed, of course.

I do a bit of VLA imaging every half decade or so, and I'm still
amazed by it. You start with what looks like a complete jumble of
numbers, mostly noise. Then you put it through a Fourier transform
(via a purpose-written program in one of the radio astronomy suites
AIPS or CASA), and "by magic" a real image appears. Of course there
are some things you have to look after (interference removal,
calibration, cleaning), but the process is straightforward in
principle.

I can't resist mentioning what might be my favorite astronomical
image of all time:
http://images.nrao.edu/AGN/Radio_Galaxies/261

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123 swil...@cfa.harvard.edu
Cambridge, MA 02138 USA

[Martin Brown]

On 02/04/2012 21:36, Steve Willner wrote:
> In article<JgDdr.25735$cd7....@newsfe06.iad>,
> Martin Brown<|||newspam|||@nezumi.demon.co.uk> writes:
>> Radio telescopes of the steerable big dish type like Jodrell Bank and
>> others like it use a technique called basket weaving which is in effect
>> measuring signal as a function along a raster grid in a particular way
>> that helps eliminate systematic errors. Typically there are a pair of
>> receivers and a difference signal recorded.
>>
>> http://ned.ipac.caltech.edu/level5/March01/Andernach/Ander2.html
>
> A few radio telescopes are now equipped with focal plane arrays, as
> another poster mentioned. I believe the one at Arecibo has seven
> receivers, for example. That speeds up mapping speed but doesn't
> create much of an image in a single pointing.

There isn't usually a lot of space for receivers at the back end
although I guess miniaturisation has advanced a lot since my day.

>> Most radio telescopes are entirely passive devices and almost all
>> serious ones at high resolution use Earth rotation aperture synthesis to
>> measure visibility fringes over a series of baselines between a set of
>> accurately surveyed dishes and from that deduce the sky brightness. The
>> observations are the Fourier transform of the sky brightness and the
>> image is obtained by computation after the observations.
>>
>> NRAO has a reasonable basic introduction for the OP
>>
>> http://www.nrao.edu/index.php/learn/radioastronomy/radiotelescopes
>
> I did a quick Google search on "aperture synthesis interferometry"
> and didn't come up with anything I liked very much. Anyone

I agree that there isn't much between the highly mathematical derivation
based stuff and abject hand waving. This one isn't too bad of the free
stuff (I suspect pinched from Jodrell Bank):

http://www.astro.ugto.mx/cursos/RadioAstronomy/Radioastronomy-4.pdf

A moderately rigorous semi handwaving method is to analyse the 1-D
linear case of equal increment baselines of the form a, 2a, 3a ...Na and
then argue that rotating it will leave a peak at the phase centre.

Most people can see it after you add one at right angles and then the
next pair at 45 degrees. It starts to look presentable after 8, 16 etc.

The other cute simulator is the Java based Virtual Radio Interferometer
http://adass.org/adass/proceedings/adass97/mckayn.html

Not sure how easy it is for a novice to make sense of it though...
http://www.jb.man.ac.uk/vri/

> interested should try it, though. There are lots of pages on the
> subject, and maybe one of them will appeal, or maybe a combination
> will be effective. It's a hard subject to explain, though, without
> making lots of diagrams and pointing at them. The basic idea is to
> use multiple antennas observing at the same time and linked together,
> but there is (normally) only one receiver per antenna. The Very
> Large Array (VLA) with 27 antennas is probably the best known
> telescope that works this way. An extension is Very Long Baseline
> Interferometry, where the telescope linking is via extremely accurate
> clocks at each telescope site rather than physical connection.
>
> The classic "student text" is _Synthesis Imaging in Radio Astronomy_.
> I have the 1988 edition, volume 6 in the ASP Conference Series, but
> I'm pretty sure there are later versions. The basics won't have
> changed, of course.

Mine is a first edition. Kraus (2nd edn paperback) is the other must
have radio astronomers handbook - a lot of good stuff in that.

>
> I do a bit of VLA imaging every half decade or so, and I'm still
> amazed by it. You start with what looks like a complete jumble of
> numbers, mostly noise. Then you put it through a Fourier transform
> (via a purpose-written program in one of the radio astronomy suites
> AIPS or CASA), and "by magic" a real image appears. Of course there
> are some things you have to look after (interference removal,
> calibration, cleaning), but the process is straightforward in
> principle.
>
> I can't resist mentioning what might be my favorite astronomical
> image of all time:
> http://images.nrao.edu/AGN/Radio_Galaxies/261

Yes. It was very impressive to see the jet for the first time.

My favourite is Cas A. I was responsible for doing the VLA 6cm A array
observations of it in 1983 (aided and abetted by Rick Perley) which back
then was the longest time allocation on the instrument. We were
operating the VLA "off piste" to get adequate aperture plane coverage
and so did the most risky observations on the first day. Luckily White
Sands missile range was quiet so we got away with it and I didn't
overload any correlators (though it was a close run thing).

http://images.nrao.edu/Galactic_Sources/Supernova_Remnants/2

And there is a movie derived from all the Cas A images
http://homepages.spa.umn.edu/~tdelaney/cas/
(sorry about the purple and green colour scheme)

Apart from the sun these two are the brightest radio sources in the sky
and so very amenable to high dynamic range imaging. Though they do test
both the hardware and software to the very limits.

--
Regards,
Martin Brown