Pattern in image

[Guest]

I was playing around with trying to split an image into three separate colour files.  I imported the CR2 files into deepskystacker, did all the calibration and exported the intermediate files as FITS. This file is not stacked, it's just one of the lights after calibration (I think). The options for reading the Cr2 was "use bayer drizzle algorithm, (no interpolation, no debayerization)". That fits file was imported into AIP4WIN where it was split into RGB as separate files (I haven't found a better way to do that yet). The green file from that was imported into AIJ where I got the image you see below with the regular pattern across it. I can adjust the stretching to remove the pattern and it then looks like a normal star picture. Is this a feature of the way the sensor works?

 

By the way, if I zoom into one of the stars I can see the green pixels and all the other pixels are black. If I zoom into the red and blue images I can see the missing pixels.

Cheers

Steve

 

[vlaiv]

This is "intermediate" stage of debayer process and pattern you are seeing is due to nearest neighbor or bilinear interpolation used to display scaled down version of the image.

It is used specifically for bayer drizzle algorithm - when stacking only pixels that are actually present (not black) will end up in stack. Algorithm expect that different subs will be shifted with respect to this and that every output pixel will have at least one "filled" pixel in the stack in its place (or more to be averaged).

What you really want is to get color images without these black missing pixels.

[Guest]

When I imported the image into Deep Sky Stacker I selected the "Use bayer drizzle algorithm" as shown in the screen shot below. It says No interpolation or debayerization so I wonder if that effect is put in by AIP4WIN or AIJ.

I'm trying to separate out  the green part of the image for photometry which we discussed in another thread. You said at the time that it's possible to do that even with black pixels but I may have misunderstood.

 

[vlaiv]

7 minutes ago, woodblock said:

I'm trying to separate out  the green part of the image for photometry which we discussed in another thread.

Do exact the same thing you are doing now, except choose fourth option - Create super pixel. This will perform debayerization completely but it wont use interpolation.

Then do the same thing - open file in AIP4WIN split channels / save green and open in AstroImageJ