Status: offline
Registered: 08/21/2009
Posts: 57
|
We have some images that involve the CCD taking four images (I'm not sure how to word it, but essentially one .fits file contains images at the NW, NE, SW, and SE corners of the display window, and a long, thick white bar goes through the images in the middle vertically and horizontally to separate them into 4 images, either it's 4 CCDs or 1 CCD with 4 slots, I'm not sure how to word it). Is there a tool in IRAF that can just remove these divisions and give me the .fits file with the four images right up against each other? I'm really not sure how or where to search this kind of a problem.
|
Status: offline
Registered: 09/30/2005
Posts: 4040
|
The answer depends on whether this is a single 2-D image with four separate image areas, or if the four images are in different FITS extensions and it was simply by displaying it that you saw the gaps.In the latter case programs like DS9 can be used to display a mosaic MEF image and the 'white bar' is an effect of how the image is displayed in the frame. In IRAF, you can try something simple like[code:1:7eac7ccb55]cl> imstat image.fits[/code:1:7eac7ccb55]If you get a complaint about 'must provide image extension', then you have an MEF file (i.e. multi-extension FITS) and you can use a task like FXSPLIT in the FITSUTIL package to separate the extensions to new images. If you don't get a complaint, or the IMHEAD task says this is a 2-D file then you'll need to use image sections to copy out the pixel areas. For example,[code:1:7eac7ccb55]
cl> imcopy image.fits[1:1024,1:1024] q1.fits
cl> imcopy image.fits[1034:2049,1:1024] q2.fits
.... and so on
[/code:1:7eac7ccb55]Presumably for this type of image the sections would be the same for all images so you could create a script to automate this.As for combining these to remove the gaps: The ARTDATA package has a MKPATTERN task you can use to create a blank image of a certain size, you can then just IMCOPY the pixels directly using an image section, e.g.[code:1:7eac7ccb55]cl> imcopy q1.fits blank.fits[1:1024,1:1024]orcl> imcopy image.fits[1034:2049,1:1024] blank[1025:2048,1:1024]
[/code:1:7eac7ccb55]Likewise, you can copy within the image to remove the gaps and then trim the excess, e.g.[code:1:7eac7ccb55]cl> imcopy image.fits[1034:2049,1:1024] image[1025:2048,1:1024][/code:1:7eac7ccb55]There isn't a task to do this automatically since this isn't really a standard format. If you'd care to post the image somewhere I'll have a look to see if there are other sugestions for you.-Mike
|
Status: offline
Registered: 11/11/2005
Posts: 728
|
Hi,I think you are talking about what is called "quadformat". Some observatories, including NOAO, sometimes used multiple amplifiers to readout a CCD and store each amplifier as a quadrant in a "flat image" (as opposed to extensions in a multiextension format). The overscan for each amplifier is placed next to the piece being read out which often means it ends up in the middle causing what appears as a stripe.Calibrating such a format is done with the quadred package. You can typecl> help quadred
cl> help quadformatto get more information.Yours,
Frank Valdes
|