PDA

View Full Version : Aperture Synthesis...how it work?



peteshimmon
2005-Nov-09, 04:03 PM
Can anyone provide a quick tutorial about what exactly happens when a
radio map is being generated? I used to think the multiple dishes looked
at a grid of points. Now I suppose each dish setting gives results for a
day that are combined into a picture after several days. So the image
generated must be a circle or ellipse that is cropped for for the normal
rectangular picture published. Dont need to know about maximum entropy!

ngc3314
2005-Nov-09, 04:47 PM
Can anyone provide a quick tutorial about what exactly happens when a
radio map is being generated? I used to think the multiple dishes looked
at a grid of points. Now I suppose each dish setting gives results for a
day that are combined into a picture after several days. So the image
generated must be a circle or ellipse that is cropped for for the normal
rectangular picture published. Dont need to know about maximum entropy!

Short form (I hope): Each measurement gives the correlated intensity between two antennae at a particular baseline (separation and direction), which measures on sine-wave component of the two-dimensional Fourier transform of the brightness distribution on the sky. It's easiest to think of this as the brightness of the radio sky within the beam limits set by the telescope aperture, multipled by a sine-wave pattern whose direction is that projected between the antennae and whose fineness of spacing is inversely proportional to the projected spacing between them (and whose phase is determined by where your electonic correlation is set to zero for). As you add more antennae, you get more measurements (the VLA does 351 at a time if all receivers are working). Earth rotation then sweeps each antenna pair along a projected ellipse, giving more complete sampling of the Fourier transform (i.e. more and more different sine-wave multiplications). As more measurements build up, the fidelity of the reconstruction compared to the image improves. (There was a whole theoretical development on just how good it would be, especially in the 1950s and ealy 1960s - you might want to check out Bracewell's book on Fourier transforms and their applications).

The folks at the Australia Telescope have produced a Java applet (http://www.narrabri.atnf.csiro.au/astronomy/vri/guide.html) to demonstrate how this works.

peteshimmon
2005-Nov-10, 01:33 AM
Thanks! As a lay person here, I think of it
like a line sweeping round on a radar PPI
display with each rotation taking a day. So
the dishes keep pointing to a fixed position
in the middle of the region of interest. I
suppose its a circular image formed at the
pole and eliptical elsewhere. Its always
rectangular pictures published. BTW the various
installations in the World must have surveyed
the whole sky by now.