Calibration Tools

[Harry page]

Hi Juan

I know this is in your mind, but I thought I would tell you what I would like

Me and others are using OSC camera and with this in mind I would like it to do

1) Debayer ( If needed)
2) Apply Flat ( Tool needed to create flat for OSC ie box car filter)
3) Application of darks and bias ( Again if needed )
4) Defect map application ( again tool to create one )

And ideally integration with the alignment and integration tool to create full automation

Just something for you to chew on as I know you don't have anything to do

Regards Harry

[Cheyenne]

I would like to chime in on adding DSLR support as well ...

[Nocturnal]

Hi Juan

I know this is in your mind, but I thought I would tell you what I would like

Me and others are using OSC camera and with this in mind I would like it to do

1) Debayer ( If needed)
2) Apply Flat ( Tool needed to create flat for OSC ie box car filter)
3) Application of darks and bias
4) Defect map application ( again tool to create one )

And ideally integration with the alignment and integration tool to create full automation

Just something for you to chew on as I know you don't have anything to do

Regards Harry

I hope you're not proposing those steps in order They should be applied from step 4 to 1 in reverse order. In the case of step 2 there is no need to do a 1:2 reduction. Instead flats should be applied on each color channel separately the way HAIP describes and DSS does. Note that this is done before debayering so a flattened image still contains a bayer matrix.

[Harry page]

Hi Sander

I should have listed them in order not as they came into my head

As you know the art of a OSC flat is a dark one

The only way I could decent results using AA4 was to remove all colour ( extract a LUM ) and I believe that the box car filter removes the bayer matrix ( Correct me if I am wrong)
But I am more than willing to listen to any ideas on this

Regards harry

[Niall Saunders]

Hi all,

The art of using OSC Flats is neither 'dark' nor 'mysterious' - assuming that you are working with RAW data (i.e. data that still contains the "CFA", in other words data that still appears 'greyscale')

ALL of your calibration steps should be using this RAW data. Lights, Darks, Flats and FlatDarks (and BiasOffsets, if you 'have' to use them) should all be 'greyscale', and should all look 'blocky' when you zoom in.

It is only AFTER you have 'calibrated' your Lights that you should be considering deBayering your (still RAW) Lights into colour frames, ready for aligning and stacking.

So, just take your Darks, Flats and FlatDarks in exactly the same way as you acquire your Lights,just bearing in mind that :

Darks should be at the same CCD Temperature as your Lights, and should have the same Exposure Time - and you should have enough of them to be 'statistically viable' (5 minimium ??)

Flats should be taken with an Exposure Time that gives you a 'Median' ADU value of around 1/2 of the maximum ADU value supported by your imager (1/2 * 65535 for a 16-bit imager, for example). This is a 'guidance level' only - you should research your imager and find out what others are recommending.

Then you would take time and temperature matched FlatDarks - and would take enough of both to be statistically useable. Your Flats/FlatDarks do NOT have to match the ExpTime or Temp of your Lights/Darks in ANY way, shape, or form.

HTH

[Cheyenne]

Here is a good "paper" detailing the whole calibration process (math included).  It's targeted towards a spectrograph, but goes into great detail the process of calibrating images.

Andreas Observatory Spectrograph Manual

[mmirot]

Yes
Time for a full blown calibration module that does these functions.
It would be nice it you could just add a set of files to the image container then have it align & intergrate after calibration.  I am sure Juan has this on his list.


Max