Wide FOV - ANSVR

[David Kemp]

Hello,

Does anyone have experience in getting wide field-of-view (150 deg, Gnomonic projection) star lists to work with ANSVR? It seems ANSVR only supports scales up to 30 degrees, but I was wondering if there is a way to download the wide-field plates and solve without requiring an internet connection to Astrometry.net.

Some background - this is for an application where I'll likely only be able to image stars of magnitude 3 or lower. Star lists would be on the order of 20-40 stars. I can get Astrometry.net solutions if I center and recalculate projections on the images for sections like Ursa Major or Cassiopeia, but I'd like to utilize the entire list of image centroids if possible (centered on zenith), and I'd like it to work in ANSVR.

I've been modelling "perfect" images with Cartes Du Ciel, then measuring centroids and inputting star lists into ANSVR, but I can't figure out how to get the plate solver to work for such large scales.

Thank you to anyone who can provide tips/hints.

[Dustin Lang]

Hi,

We never really pushed Astrometry.net to the large-scale limit.  It internally assumes a TAN projection, very fundamentally.  If you can pre-distort your detected source positions to match that projection, then you might have some luck, but I have never tried that.

cheers,

--dustin

--
You received this message because you are subscribed to the Google Groups "astrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to astrometry+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/194502ed-44c3-48b8-8231-8cb88b1a764fn%40googlegroups.com.

[Dustin Lang]

(ps, the solve-field code does have a --predistort option to undo a known SIP polynomial distortion, but the web service doesn't currently have a web of putting one of those in.)

[David Kemp]

Hi Dustin,

The imagery (or centroid lists) that I am using already have the tangential projection accounted for. Is there a practical scale limit to solver? I noticed "ultra-wide field" options of 10-180 degrees, but I have yet to make anything beyond 75 degrees work.

Thank you,

David Kemp

[Dustin Lang]

Okay, that's interesting.  Our largest-scale standard index files (scale 19) have quadrangles that are ~23 to ~33 degrees across.  So maybe you don't have enough stars to create a lot of features of that scale.

But also on the projection front, I haven't thought about this very much, but I'm not sure that we correctly deal with the way shapes get distorted away from the center of TAN images.  We try to match the shapes in pixel space to the shapes of known stars, assuming a TAN projection *centered on those stars*.  Could it be that we would have to correct for the image-space shapes based on their radius (but that would be a radius in *degrees*, hence requiring knowing the image scale...).

cheers,

--dustin

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/31b358e0-905f-415b-99ca-117d4ab01069n%40googlegroups.com.

[Eric SIBERT]

For (very)wide field, you may crop the center of the image. For ~100°, I
take the half center. For fisheye, I use the quarter.

Eric

> --

[David Kemp]

Hi Dustin,

In terms of image scale and separation of objects in degrees, I have a system that has a projection mapping of point-source centroids to input-angles, so going from a tangential projection to a solid-angle projection would not be difficult. However, even knowing the image scale and the solid-angle separation of objects in the image - is there a way to "inject" that information into the algorithm?

-David Kemp

[Dustin Lang]

Currently, no, there's no way to inject that.

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/d90b4cff-2cf4-4eac-981b-edfa18637599n%40googlegroups.com.