Welcome to iraf.net Thursday, March 28 2024 @ 05:52 PM GMT
mdcleir |
08/31/2017 08:59AM (Read 1417 times)
|
|
|
Status: offline
Registered: 08/30/2017
Posts: 1
|
Hi,
I am using the msccmatch tool to match the WCS of different exposures.
I want to use this task interactively in order to check if the fitting does a good job.
When I run the task, the first plot indeed pops up in the xgterm.
But, if I want to delete certain data points with the [d] key, random points are selected, and not those nearest to my cursor.
This is pretty annoying because it is impossible to select the data points that I want.
I found other discussions about this, but none of the solutions was really helping for me.
Thank you in advance,
Marjorie
|
|
|
|
valdes |
09/08/2017 03:51PM
|
|
|
Status: offline
Registered: 11/11/2005
Posts: 728
|
I have a couple of questions before we investigate whether it is a system problem.
When you start msccmatch you will see a plot that represents the positions of the sources. This is mostly useful to see if you are missing the corners or otherwise have a poor coverage. So you should type 'x' to see fitting residuals as a function of the ra coordinate. Does the 'x' key respond and show you a different plot from the first? Deleting points is usually done in one of the residual plots (x, y, r, t) rather than the initial positional plot. Are you deleting from a residual plot or the first plot spatial plot? Is you plot window roughly square or highly elongated?
From the command line in the xgterm, say after you quit msccmatch, type '=gcur'. This will return you to the graph window and bring up the cursor. Put the cursor near a point and type any key. This will show you the cursor return. What does it say?
Concerning the question about "trends of residuals" I mean whether there appear to be wiggles in the mean scatter or if the scatter looks random and centered on zero everywhere along the independent variable axis. For instance with very low orders (say 2) you might see a clear quadratic shape to the scatter of points. This indicates that a slightly higher order will take out this trend.
Cheers,
Frank
|
|
|
|
| |
|
Content generated in: 0.05 seconds |
|