Starlight Live Lodestar-C colour decoding improvements preview

[Paul81]

Hi All,

I have been working on improving the tricky colour decoding of the Lodestar-C CYGM bayer sensor. I have switched to a different method of converting the CYGM colorspace to RGB, so I thought I'd quickly post some comparison pictures.

Here is a shot of a test card using the equations from the current released build:

And here is the result from the modified code:

In terms of real DSOs, here is a shot of M27 from the current release code:

And here from the modified code:

In these comparisons I have used the same display processing settings, so the differences are purely a function of the new code.

I have also been working on an automatic defective pixel removal, with the following results (these images do not have darks applied):

Defective pixel removal disabled:

Defective pixel removal enabled:

I'd still recommend darks to get rid of other noise, but the defective pixels algorithm seems to work reasonably well, needs a few tweaks but its progress (it introduces some artefacts around bright stars which I want to get rid of)!

Depending on timescales for V3.0, I may do a V2.2 update with the new CYGM code, but defective pixel removal will be a V3.0 feature.

Paul

[Bukko]

Nice, Paul,

Having just purchased a new Ultrastar last week (Mono), I am hoping to get to try the new software soon. I have started a little bit of outreach with some local kids and I thought this will be a brilliant way of showing them images appearing as they watch.

Keep up the good work !!

Gordon.

[DoctorD]

Hi Paul

Colour decoding looks much better and the defective pixel removal works well - I recognise that Test Card

If you'd like more test data I'd be happy to provide FIT files from my Lodestar-C.

CS

 

Paul

[nytecam]

Nice work Paul - the new colour saturation is much better.  Bit surprised with the number of hot-pixels removed by the new ADPR routine !

Nytecam