Astro Image J (AIJ) strangeness

[Guest]

I start with CR2 files from my canon camera, calibrate them using deep sky stacker which generates for me a FITS file for each image. The fits files contain 3 layers - one for each colour. I've confirmed this by examining the FITS headers.  But there is a peculiarity with the way AIJ loads these files.

If I import a single fits file AIJ says number of images = 3. If I then carry out the photometry AIJ gives me 3 values for the variable star magnitude. These are labelled slice 1,  2 and 3 and refer to the same filename. I assume that is one magnitude for each colour R, G and B.

If I import two fits files AIJ says number of images = 4. If I then carry out the photometry AIJ gives me 2 magnitudes for the variable star which are labelled slice 1 and 2 corresponding to the two input files.

If I import 28 files AIJ says number of images = 28 (which is correct). If I then carry out the photometry AIJ gives me 28 values of magnitude which are numbered as slices 1-28 corresponding to my input file.

Does anyone know what's going on here?

In the case of importing 2 images I've tried changing the number of images to 6 (2x3) but it still only gives me 2 values of magnitude.

What I want is for AIJ to work out the magnitude separately for each colour so if I import 2 mages I want it to give me 3 values of magnitude.

 

 

[matt_baker]

My best guess would be to extract the green from every single one using a program then using AIJ with those?

I saw your other post, maybe I can look at your csv just after your multi aperture and see what I can come up with?

Also I could look at all of your data myself if you upload the files?

Edited December 18, 2020 by matt_baker

[vlaiv]

Yes, ImageJ has very strange handling of color files and it's best not to treat them as color images - but rather as separate channel images.

It has concept of :

- stack

- hyperstack

- color stack

- composite image

And I have no idea how it automatically transforms between these.

If you for example want to make RGB image from channels, you need to make stack of 3 images - R, G and B.

Then convert that to 16 bit mode (I do all my processing in 32bit float point - that way you won't introduce any rounding errors that are significant) it uses display brightness / contrast controls - that are otherwise non destructive and are just for display purposes.

Once you are in either 8bit or 16 bit mode - then you can use Stack to RGB command.

In any case, workflow that I would recommend would be:

1. use software like FitsWork to convert DSLR raw files to fits. Resulting fits will be 16 bit format

2. use ImageJ to convert those 16bit fits to 32bits and then calibrate your images there (make stacks, subtract, divide - regular calibration workflow)

3. use debayer plugin to debayer images. One that I've found online works for 16bit images, but I have few that work with 32bit images - so if you need them just let me know and I'll attach them here (you can either do interpolation or split bayer matrix - which ever you like)

3.1 alternatively if you don't want to deal with debayering - you can just bin your raw images 2x2 - that will add 4 bayer pixels together and you'll get photometric result over whole spectrum.

4. open resulting images in AstroImageJ and measure things ....

If you need any help with ImageJ / Fiji (how to calibrate or debayer plugin or whatever) - let me know, I'll be happy to help.