coordinate ambiguity in ccmap

Frank Winkler wrote on Jan 23, 2012

I am trying to refine the WCS on images for which I have precise pixel (logical) positions and RA, Dec for only two stars. (I'm working with Chandra X-ray data with only a few point sources.) Two stars are sufficient to get a plate solution using rscale in ccmap, ... except that there is an ambiguity in handedness; i.e., there are two possible solutions that are flipped about the axis running through these two reference stars. Unfortunately, I seem to be getting the one opposite to the correct one. Is there some way to ensure that I get the correct solution?
Thanks!
Frank Winkler

---

Francisco Valdes wrote on Jan 23, 2012

Hi Frank,

I'm sorry I did not catch this post sooner. I'm surprised that ccmap even works with two few stars to resolve this ambiguity. What I would do is add a dummy point to force the right "handedness". Is this still a problem? Let me know some details, like sending me the input file and parameters, and I can look at it.

Yours,
Frank Valdes

---

Frank Winkler wrote on Jan 23, 2012

Hi Frank,
Adding a dummy point is precisely the approach I took on my own--starting with something close and then adjusting coords of the dummy until x and y residuals become negligible. So this works OK; I was just wondering if there was a shortcut, but apparently not.

But it is curious that with 2 points ccmap DOES give a solution, it's just always the one with the opposite handedness.
Frank/Paco

---

Last post on Jan 23, 2012