Problem when solving u-band images.

[Jamin Welch]

Hello all,

    I am trying to get WCS solutions to several images.  I am doing this with the off-line version of the astrometry.net software.

The command I have been using is "solve-field --scale-units degwidth --scale-low .1 --scale-high .3 image.fits --uniformize 0  -O"

My images are ~13.7 arc-minutes wide and tall. and I have all the 202 and 204 indexes but with the given scale it uses the 202 index images exclusively.

In general it will solve other filter images just fine but when It approaches a u-band image for some reason it will not even select stars in the field that seem usefully bright, instead it will select dimmer points in the background. Does anyone have any ideas as to why that is?

I have attached one of the red .fits images and one of the ultraviolet .fits images that I am attempting to solve as well as there respective objs-png files.

Most of the u-band images simply will not solve and if they do the solutions are dead wrong.

The images are here --> https://drive.google.com/open?id=1jiO8VRqPZ4Ih6KPCEtSPykmOtsXZ831l

[walde...@gmail.com]

converted ultraviolet.png to jpg.  solved fine when uploaded.  try comparing.

[Jamin Welch]

The issue there is I need the actual data from the fits image. I have a large set so it isn't feasible to have to upload every image and then just download the correct wcs file and apply it. But I appreciate your input.

[Dustin Lang]

For the source detection, try using "--downsample 2", which is what the web version does and tends to give better results.

However, looking at the sources detected in red vs ultraviolet on the web version, the source detections look about equally good, so I'm pretty disappointed that the UV one doesn't work.  (the red solves after looking at only 13 sources)

There is a reason to expect red to work better, FWIW: the 200-series index files are built using the "red" emulsion to sort by brightness, so there is a mismatch between which stars are considered 'bright'.  But I would expect it to work a little worse, not fail entirely.

--dstn