iraf.net

Hi!I built a mosaic, which consists of several exposures. Since those exposures were not taken on the same night, all of them have slightly different background levels. Does Iraf provide a task to flatten the background levels of all images to the same value?
Or can you suggest another software which is able to do so?Thanks.

There are several ways this is done. If the backgrounds are flat (or the same shape but with a constant offset) then this is generally handled during combine with IMCOMBINE. The imcombine zero scaling parameter can define the values to come from a header keyword for each image or from a file containing the scaling values.Another way this is done when the backgrounds have variable gradients or shapes is by subtracting the background from each exposure prior to combining. This might be done with something like IMSURFIT where a surface is fit and subtracted. Of course you need to mask out the sources in some way. There is a task in mscred called MSCSKYSUB that might help since it can use masks. A mask of the sources can be created with something like proto.objmask. Note the trivial subtraction of a constant that you determine with something like IMEXAM would be done with IMARITH.Often when combining there are actually two effects to handle, the transparency which is a multiplicative scaling and a sky brightness which is an additive one. It may be your background levels are varying also because of transparency (or exposure time). Transparency variations are normally determined from making the source counts in common objects come out the same. The task mscimatch is an empircal way to measure the scale and offset values. Or you can use a photometry package to measure the counts and sky. Then you would use imcombine with one or both of the zero and scale parameters.So this problem is one of the most common astronomical image processing operations and you have to decide what is right for your data after which there are plenty of tasks to either estimate the matching or to apply it before or during stacking.Frank Valdes

Hi!Thank you. Im very new to iraf, so I need much mir guiding Speaking of imcombine... I saw that imcombine can use the header of my fits to not even combine the images, but also make mosaics of them. Only problem is, that areas that overlap have some sort of frame around them. Is there a way to get rid of them easily?thanks.btw: Im using SDSS Images, and in those fits headers, there are sky values given in DN/pixel. Can I simply use them as a zero offset by choosing !Sky as a zero value?

I don't understand what you mean about the "frame". As for !SKY, it is likely that that would work. Or it might require changing the values to be the negative of the value.To help you you could send me the parameters you are using (the output of "lpar imcombine") and the header listings of a few images (the output of "imhead <images> l+".Frank

Thanks for the parameter and header information. I don't think I've ever looked at the public SDSS data headers. If I understand the headers, the data are in ADU (i.e. data numbers). Therefore, you need to use the SKY keyword. However, it has to be the negative of the value for imcombine. You can add a keyword with:[code:1:c541c1c623]
hedit <images> SKY1 "(-SKY)" add+ verify- show-
[/code:1:c541c1c623]As far as any "frame" that you see, this might be something in the data.I hope this helps.Yours,
Frank[quote:c541c1c623]
[parameters and header listings deleted]I hope its not to much... as you can see, the exposure time is always the same.With the above configuration, columns of images seem to have the same background level. I guess this is because Sloan takes images by letting the sky drift across the ccds, and different columns, or stripes as they call it, are aquired at differend days.
[/quote:c541c1c623]

Hi!Sorry, somehow I didnt get a notification about your reply.
Thanks for the effort!But now I do my mosaics with "Montage". It has a task specialized for SDSS data. its quite good.as for the "non-flat" background... is it possible to flatten it without having any flatfield images? In my case, I only have those from SDSS, and they do not offer any flatfields.
Is there some kind of tutorial on background extraction? I checked those tasks you mentioned, but they seem confusing. The best thing I managed so far, is to use SExtractor's background estimation to extract the background, but often it introduces severe artifacts... blocks or large stripes that are black or white...