Longitude range selection
Page 1 of 1

Author:  deil [ Fri Aug 27, 2010 5:10 pm ]
Post subject:  Longitude range selection


I noticed that for skymap_format = 1 the Galactic center is always at the side of the image.
Is it possible to have it at the image center and GLON increasing to the left, as is common for all-sky images?

And would it be possible to add dlong=0.1 and dlat=0.1 as valid parameters?
I would like to get a high-resolution view of the Galactic plane.


Author:  strong [ Sat Aug 28, 2010 11:26 pm ]
Post subject:  Re: Longitude range selection

Since the format is FITS, the start of the longitude axis it arbritrary and was chosen as l=0 for simplicity.
Where the image is centred depends on the plotting software. If displayed with e.g. fv indeed the GC is at the edge, but more advanced display tools can centre it anywhere.

Apart from that, GALPROP has the parameters long_min, long_max which default to 0 and 360 (actually 0.5 and 359.5 since they are bin-centres for 1.0 deg bins). You could try -180 to 180 (or -179.5 to +179.5), but for some reason this is not allowed in the current menu. When enabled it should work although I have not tested it with the current version.
Similarly dlong negative should work, but it not allowed in current menu, and this needs changing.

(There should be an option to allow the user to enter any parameters overriding validations, but I don't know if that is available yet).

re dlong =0.1 and dlat=0.1, those parameters are valid: Webrun does not have them but it should be changed to allow any value.

Note also the Healpix output option, which has the advantage of being uniform on the sky, essential if you want to study high latitudes. Use a high order like 8 or 9 to get a hi-res view of the sky, but beware of very large output datasets. Again this in not in the drop-down list but should be.

Author:  deil [ Sun Aug 29, 2010 6:22 am ]
Post subject:  Re: Longitude range selection

The possibility to run glon_min = -180, glon_max = +180, dlong = -0.1 with WebRun is exactly what I need.

One more suggestion: I found some GALDEF files on the web, corresponding to publications.
It would be very useful to be able to upload a GALDEF file, and then select it from the list "Import configuration from" on the configure & submit page.


Author:  strong [ Sun Aug 29, 2010 7:38 am ]
Post subject:  Re: Longitude range selection

in fact you want long_min=180, long_max=-180, d_long=-0.1

I will check that this actually works and then ask Andrey to widen the options here.

re published galdef files: good idea, something else for our to-do list.

Author:  strong [ Mon Aug 30, 2010 12:47 am ]
Post subject:  Re: Longitude range selection

I tried various of those things but did not get the required effect for centring or reversing. So you are stuck with l=0-360 and +ve d_long for now I am afraid.
Still to shift/reverse the axes is a few lines of code, I can provide it if you want.
Which imaging software are you using ? fv is quite basic, but I would hope ds9 etc would be able to do simple things like shifting or reversing the image. If you are using idl, shift and flip are simple to perform.

For the fine scale, 0.1 degrees will work, just has to be allowed in the menu. However note that the gas surveys used are limited to 0.5 degrees, so there is actually no point going finer.

Meanwhile let me recommend healpix again, which is the projection of choice of all-sky maps in astronomy nowadays.

Author:  deil [ Mon Aug 30, 2010 8:51 am ]
Post subject:  Re: Longitude range selection

Unfortunately HEALPIX is not an option for me, because I have to compare to other data in cartesian coordinates.

Concerning FITS viewers: I am using ds9 and Aladin, both of which don't have the functionality to switch left / right image halves, which in my opinion is not a common enough operation to be included in a viewer.

The following python / pyfits code snippet does the trick, though, as long as the number of pixels on the GLON axis is even:

header = pyfits.getheader(infile)
sb = pyfits.getdata(infile)
# Convert GALPROP output to what I want, i.e. the
# Galactic center at the center of the image
half_naxis = header['NAXIS1']/2
left = sb[:,:,:half_naxis]
right = sb[:,:,half_naxis:]
sb = np.dstack((right, left))
# And the GLON axis increases to the left
sb = sb[:,:,::-1]
# Now adjust header keywords to new array.
# Note that pixel 1 will go to where pixel
# NAXIS1 / 2 was.
header['CDELT1'] = -header['CDELT1']
header['CRPIX1'] += half_naxis

I'm looking forward to the possibility to set d_long < 0.25 on WebRun, so that I can get high-resolution views of the Galactic plane.


Author:  strong [ Mon Aug 30, 2010 9:06 am ]
Post subject:  Re: Longitude range selection

for the viewers, one would think that to display a FITS file in the astronomical sense independent of how the pixels are
arranged in the file, would be basic. Apparently not !
I will ask Andrey to allow d_long=0.1 but as I said this does not mean much on account of the gas surveys resolution.

Author:  deil [ Mon Aug 30, 2010 9:10 am ]
Post subject:  Re: Longitude range selection

Sorry, I had missed your comment that the resolution of the gas maps is 0.5 deg.
In that case disregard my request for the possibility to select a finer resolution in the WebRun interface.


Author:  strong [ Mon Aug 30, 2010 9:21 am ]
Post subject:  Re: Longitude range selection

in any case we should give more freedom for the user to choose other values than in the menu.

Author:  gudlaugu [ Tue Aug 31, 2010 4:43 pm ]
Post subject:  Re: Longitude range selection

There is a reason for limiting the possible values in the menu, the biggest one being that the pixel boundaries have to align with the input gas maps. Therefore, only limited values of dl are allowed. Currently, the code is also written in a way that it can't handle negative values for l nor negative dl values. Since the maps can easily be rotated in post processing, I am reluctant to add this feature to the code. But it can be done if this becomes something people really want. File a feature request in the bug tracker so this won't be forgotten.

Author:  strong [ Tue Aug 31, 2010 11:28 pm ]
Post subject:  Re: Longitude range selection

Gulli is right, it fact the map-making is tricky and has to be tuned to the input gas maps for optimal results.
Since indeed there are plenty of ways to post-process the maps, we can stay with the present options.
I had thought that at least shift or reversal was handled by GALPROP (this would not have those alignment issues) but in fact not.

Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group