Unusual source extractor pixel coordinates and rounding errors

16 views
Skip to first unread message

Serge

unread,
May 18, 2018, 2:46:41 AM5/18/18
to astrometry
If I upload a black and white image with stars replaces with single white pixels nova.astrometry.net returns an axy.fits (Stars detected in your images x,y table) that looks something like this:

...
  265.9551   79.0449    9.0074         0
  101.9551  159.9551    9.0074         0
  247.0449  225.0449    9.0074         0
  121.0449  259.0449    9.0074         0
...

I expected all pixel coordinates (columns 1,2) to all end with *.5,*.5 or all end with *.0,*.0.
It looks like they are all close to *.0,*.0 but +-0.05, which is a bit unusual and perhaps sometimes significant.

Personally I prefer the coordinate system, that has image pixel corners falling on round integers and pixel centres ending with 0.5,0.5 because that coordinate system has a  lot of natural advantages, but that's another story....

However it may be useful to let users know when they upload their own x,y tables that they must assume that the [left,top] image pixel centre falls on [1,1] and the [left,top] corner of that pixel at [0.5,0.5].
Perhaps it is already mentioned and I am blind... if so my apologies.

I know that I can simply upload my own x,y table, but I wanted to test is source extractor behaves sensibly on simple data.

Here is the full table:
  168.7822  285.7421    9.0089         0
  164.1106  279.2492    9.0089         0
  343.9551    3.2671    9.0074         0
  233.0449   19.9551    9.0074         0
  119.9551   87.0449    9.0074         0
  459.9551  227.0449    9.0074         0
  145.9551  431.0449    9.0074         0
  301.9551  439.9551    9.0074         0
  265.9551   79.0449    9.0074         0
  101.9551  159.9551    9.0074         0
  247.0449  225.0449    9.0074         0
  121.0449  259.0449    9.0074         0
  411.0449  391.9551    9.0074         0
  221.0291  477.9392    9.0074         0
   57.0449  471.0449    9.0074         0
  197.0449  115.0449    9.0074         0
   29.0449  179.0449    9.0074         0
  241.0449  245.9551    9.0074         0
  137.0449  291.0449    9.0074         0
  215.9633  467.9674    9.0074         0
  119.0449  543.9551    9.0074         0
  257.0449  141.0449    9.0074         0
  177.9551  305.9551    9.0074         0
  187.0449  513.0449    9.0074         0
   21.0449  565.9551    9.0074         0
  175.9551  169.9551    9.0074         0
  209.0449  381.9551    9.0074         0
  289.0449  533.9551    9.0074         0
   91.0449  603.0449    9.0074         0
  201.9551  173.0449    9.0074         0
  265.0429  579.9526    9.0074         0
  261.0458  567.9569    9.0074         0
  259.0449  612.4665    7.1247         0

And the link:
http://nova.astrometry.net/user_images/2116794

Best regards and thank you,
Reply all
Reply to author
Forward
0 new messages