Have image processing algiorithms changed recently? about indi-allsky HOT 9 OPEN

I did change the default Field of View used for the SQM calculations and processing. I changed the default to the central 50% of the image where it used to be 66%. If you change the SQM FoV to 66% on the Processing Tab, it should function like it used to for you.

Another workaround is creating a detection mask, which will use the mask for the borders of the calculations.

This most heavily affects the stretching algorithm.

from indi-allsky.

Following on from this and my post on the AllSky group on Facebook, I'll try the new release and report my findings here. The 'false colouring' does seem related to SCNR as disabling this does stop the colouring. I will also try setting the colour saturation setting back to 1.0, as 1.5 does seem to over-colour my stars.

from indi-allsky.

Those colors actually look very good.

I have learned quite a bit more about numpy since originally implementing the SCNR. I now know how to deal with the integer overflows.

from indi-allsky.

I've updated the software following your note and I'll see how things go, so thank you as always for the quick response.

To me the colours look too magenta, if you follow. I almost want to rebalance the colours. Maybe I will do that once things have settled down with the processing.

from indi-allsky.

Well, I'm still getting very odd colours...

I've lowered the gains in case I'm getting a lot of large-value pixels and I've got CLAHE and SCNR disabled; stretching on, but not excessively. My settings are:

Night Gain | 1800
Moon Mode Gain | 1400
CLAHE Clip Limit | off
CLAHE Grid Size | off
16-bit Contrast Enhance | off
Enable Stretching | on
Stretching Gamma | 1.5
Stretching Std Deviations | 1.0
Moon Mode Stretching | on
Daytime Stretching | off
Stretching split screen | Off
Bayer Pattern | RGGB
SCNR (green reduction) |  
Red Balance Factor | 0.6
Green Balance Factor | 0.8
Blue Balance Factor | 1.0
Auto White Balance | on
Saturation Factor | 1.5

The histogram is

Is there a setting I can use to get the raw out-of-camera FITS file so that I can see what the pixels are before processing?

from indi-allsky.

The moon is just oversaturated. Combine that with the stretching and the auto white balance being enabled, you could get the result you are seeing. The moon halo has a slight red bias, but that is not entirely unusual.

from indi-allsky.

The histogram may be a bit deceiving, too. I just remembered that the histogram is generated from a crop of the center of the image (trying to exclude things like trees on the edges).

from indi-allsky.

I'm still bemused by the incredibly strong colouring I am getting. I've reduced the gain of the camera; turned off CLAHE and have no colour saturation enabled - and I am still getting very strong colours in my images which are not present in the FITS versions of the images.

Is there a way of setting all processing settings to their defaults in case I've missed a setting? I can then start my processing set up from scratch again.

from indi-allsky.

There is a new feature in the latest release where you can test processing settings. If you have a FITS version saved of an image (in the database), you can select FITS Processing in the image viewer (at the bottom) to bring up a page where you can test different processing settings. That might help you track it down.

from indi-allsky.