gnuastro-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gnuastro-devel] [task #14368] No fits correction for modular warpings


From: Mohammad Akhlaghi
Subject: [gnuastro-devel] [task #14368] No fits correction for modular warpings
Date: Thu, 9 Feb 2017 20:28:58 -0500 (EST)
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:51.0) Gecko/20100101 Firefox/51.0

URL:
  <http://savannah.gnu.org/task/?14368>

                 Summary: No fits correction for modular warpings
                 Project: GNU Astronomy Utilities
            Submitted by: makhlaghi
            Submitted on: Fri 10 Feb 2017 10:28:57 AM JST
         Should Start On: Fri 10 Feb 2017 12:00:00 AM JST
   Should be Finished on: Fri 10 Feb 2017 12:00:00 AM JST
                Category: ImageWarp
                Priority: 5 - Normal
              Item Group: Enhancement
                  Status: In Progress
                 Privacy: Public
        Percent Complete: 0%
             Assigned to: makhlaghi
             Open/Closed: Open
         Discussion Lock: Any
                  Effort: 0.00

    _______________________________________________________

Details:

The FITS standard defines the center of the first pixel to have a coordinate
value of 1.0 along every dimension. As a result, the center of coordinates is
defined to be the half a pixel outside the input image.

As a result, when using the higher-level modular warpings (for example
`--align', `--rotate', `--scale', `--translate', or `--project') ImageWarp
will currently shift the center of coordinates by (+0.5, +0.5) so the bottom
left corner of the first pixel will be at the coordinates of 0.0.

By by preferentially correcting one side of the image, this behavior can cause
very confusing behavior (for example task #14363). Currently ImageWarp has the
`--nofitscorrect' option to stop this correction on modular warpings. 

So I am thinking that if we remove this higher-level correction so the edge
pixels on both sides of the image will be similarly affected by the warping.
This can help avoid confusions about how ImageWarp operates. 

If users like to have preferential behavior on one side of the image (put the
center of coordinates at the bottom left corner of the first pixel), they we
can reverse the `--nofitscorrect' option, with an option like
`--centeroncorner'. Alternatively, they can use the `--translate' (`-t')
option before an after their desired transformation, like below:


$ astimgwarp -t0.5 --scale=2 -t-0.5 img.fits


As part of task #14245, I have just started working on ImageWarp, so I wanted
to see what your comments on this are.




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/task/?14368>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]