Tiny Fields

[Judy Schmidt]

Astrometry is still kicking my butt after all these years, and I still have not managed even one solve. I have altered my strategy! Tonight I got a good source extraction with almost all sources detected correctly, and only a very small number missed... however, I am afraid my FOV is too tiny. This is the same FOV as Hubble's anniversary Horsehead Nebula, and I was trying to find a way to get it into World Wide Telescope. My strategy is to align a DSS visible light image to the infrared image so that the right kinds of sources are visible, and then later swap the image out to the real image.

I was hoping that this would be large enough because it is a 3x3 mosaic so it should be better off than a single frame, but alas. I am afraid it is not so.

Without further ado, here is Failhorse. Is Failhorse doomed?

http://nova.astrometry.net/user_images/1992112#original

[Judy Schmidt]

Through a related line of inquiry elsewhere, I have incidentally confirmed that this failure is typical, and that I'm probably not making a mistake.

[Dustin Lang]

Hi Judy,

The nova.astrometry.net version uses the USNO-B1-based index files, and I'm guessing that USNO-B1 is a mess in a complicated place like the Orion nebula.  That, plus the small FOV, makes it pretty tough.

I just tried creating a custom index from Gaia in your region, and got a solution.  I'm attaching it here.

cheers,
--dustin

[Dustin Lang]

PS, I like how the solving quad looks like a unicorn horn.

[Judy Schmidt]

Ok, what's really neat about this is that, if I am understanding correctly, those green circles are all sources from Gaia, and those are actually present in the Hubble infrared image, so that's really interesting to me. Previously the infrared data had to be substituted with DSS data because all those extra sources were just useless. See: http://geckzilla.com/astro/HST_Horsehead_AVM.jpg

I ended up getting WCS data from the original FITS files, converting them to AVM data, and then manually adding the AVM headers to that jpg I linked to with some tweaks I made by hand. Not ideal, obviously, but it works in a pinch. This, as I have discovered, has been done in the past with many Hubble images, though it is so tedious that it has seemingly been abandoned for the past few years. So we have websites like http://astropix.ipac.caltech.edu/ that offer these images for outreach purposes, and they try to include AVM headers, but some of these fantastic Hubble images such as the Horsehead, the anniversary version of the Eagle Nebula, etc. do not have any coordinates.

You've given me several files that I don't know how to utilize, unfortunately. I'm really sorry I don't know how, because I'd like to copy the solution to my Hubble image headers, and see how it compares to the one I did manually. I'm sure the solution is much better than what I did, which was too tedious a process for me to get pixel perfect. I wonder if what you've done can be further automated so that Hubble images from ESA and STScI can finally be given the missing AVM coordinates more easily? I wonder if my own personal collection of Hubble processing could also benefit. 

As always, you have done so much more than I expected. And I know I don't have to say it, but it is totally ok if you don't have any more help to offer. I know you must be terribly busy, and I almost wonder if you have clones doing all the work. I feel like I've asked too much lately.

The unicorn horn is great. 

[Dustin Lang]

Hi Judy,

WOW, that infrared horsehead image is spectacular!  I'm sure I have seen it before, but WOW!

One of these years we will actually connect all the dots between FITS and JPEG :)

What tools do you use to plug AVM tags into your images?  I think I was waiting for virtualastronomy.org to produce the (seemingly simple) tool to take a WCS header and a JPEG image and plug in the AVM headers, but that never seemed to materialize, and now the web tool is 403 forbidden...
https://www.virtualastronomy.org/avmwebtool/

Ahh, but then 10 seconds on google and I see
https://github.com/astrofrog/pyavm
which could be interesting!

cheers,
--dustin

[Dustin Lang]

PS, we always intended the Astrometry.net tools to be useful for a wide variety of folks working with astronomical images, and you are definitely in that set but there's still some mismatch between what we've got and what you need, so I find your reports super helpful in figuring out what we should do better or different!

[Judy Schmidt]

There are a couple of different ways I learned about.

An online tool here provides a webform for easy editing of the headers, but requires one to upload the file, edit the headers, and then download the updated file.

http://avm.ipac.caltech.edu/

Far more preferable for me, however, was to have the astronomy AVM metadata editable directly with Adobe products. The link to the files required for that is now available and working at the IPAC AVM website. (Adobe keeps changing the way those extensions work, so the files keep having to be updated, and web searches are full of dead ends and broken links)

http://avm.ipac.caltech.edu/page/7

This is, of course, not an ideal solution for everyone, since Adobe software is pricey, and not everyone uses it. Too bad there is no freebie standalone editor. It makes sense that it works so well with Adobe products though, since Adobe created XMP.

Judy

[Judy Schmidt]

I am very relieved to hear that I am perhaps less of a nuisance than I imagine myself to be.