ccdproc
ccdproc copied to clipboard
astropy
Implement cross-talk correction
My feeling is that this isn't needed in v0.1, but it would be nice additional functionality to add. It will be important for ccds read out from multiple amplifies. However, it is something that can be done just through slicing and arithmetic. In addition, it might be difficult to make this general enough but still useful as different configurations may behave differently.
I'm not sure if that helps: For our data reduction of data taken with the OMEGACAM camera at the VST, we have written some code for cross-talk (CT) correction. What we here discovered is that the CT coefficients changes even during the night, but they can not always be estimated and then a default value is needed. So how we did it: Knowing from which chip the flux is going to which chip, we
- detected all objects on the chip where the flux was added
- inverted the image where the flux was stolen so these areas become positive and detected those objects there
- checked for identical positions in both and then calcuating the ratio.
Doing this on exposure base results sometimes in completely different values, even from exposures right after each other. So it should be included how the correction value should be calculated.
The actual correction is then simply dividing / multiplying both images with the corresponding values. Furthermore, one chip can steal flux from another chip and at the same time flux is stolen from it from a third chip.
Sounds like a very promising method. Do you have code to make a pull request and/or a link to the code if we wanted to include this?
I'll ask the people who wrote it. btw I forgot to write that it's a bash script and we use SExtractor. But still, it might help as guideline...
Being able to use it as a guideline (and perhaps a cross-check?) would be great, @dklaes
I just talked to the authors, here is the code (I hope it's well enough commented so it makes sense): https://gist.github.com/dklaes/74a43f6d57f75a496c14 https://gist.github.com/dklaes/7973d85d5712f2b959bc https://gist.github.com/dklaes/840a413c57ef411d9777