Overscan processing or removal

@deisenlord,

How do you save the overscan? Do you select it as an option in the capture application?

Or is it special metadata in the fits headers? Probably I need to see exactly how it works for your camera to be able to implement this.

Mabula

The overscan is always in my FITS file from the Indilib FLI driver which uses the Finger Lakes Instruments Linux driver.   Unfortunately there aren't any keys in the file to describe the overscan region.  I've contributed to this driver for Indi and could reach out to Finger Lakes to see if they know of any "standard" format for adding overscan keys to FITS files.   

I found I could use your  crop tool to remove the overscan , in my case its a 40 pixel wide strip across the top and a 40 pixel wide strip down the left side. I could just set the offset to 40,40 to crop it out.   So I have a workaround that pretty easy to apply.

If you'd be interested in processing the overscan I could reach out to Finger Lakes to get their input on adding FITS header info describing the overscan and look at adding it to the Indilib FLI driver.    Only if you think this is worth it for APP long term as I'm good for now.

Best,

David

Hi David @deisenlord,

Excellent, yes the offset crop tool is the answer to be able to remove the overscan from your calibrated files before starting with registration etc...

Yes, if they camera maker/capture software maker adds overscan information in the metadata/fits header, it will be very easy to fix in APP 😉

Without the information in the metadata, it will become very cumbersome quickly.

Another way would be to provide the overscan offsets manually in the calibration menu. So then, after calibration APP can automatically remove the overscan based on the supplied offsets. I can implement that fairly quickly I think. Would that be a good option/alternative?

Mabula

This is the way PI handles it in their Batch Processing Script, a set of fields where you can describe the overscan regions. The documentation is non-existent so it took a little messing around to get it specified correctly.  I researched FITS keys and it seems the BIASEC and  TRIMSEC are commonly used in research environments, but certainly not always, example.

BIASSEC=   [2054:2101,10:2045]) Overscan strip image section
TRIMSEC=     [51:2002,51:2002]) Trim data section

I'm pretty sure this isn't used enough to warrant implementing though, I think your idea of adding options in the calibration menu would be more widely applicable and a better solution.

Hi David @deisenlord,

I will study the FITS specification for these tags.

Indeed, providing overscan offsets in 2) calibrate at least should work always then.

This is on my ToDo list to implement.

Thanks for bringing this to my attention 😉

Mabula