So I considered having an integration option as a last possibility and may still add that option, but doing that does require more memory and cpu since it would require creating a whole new buffer and iterating over all the pixels in the image to calculate all new pixels. This sounds like no big deal unless you consider the size of images and the numbers of pixels and the numbers of times the calculations would need to be done. Also integrating all three channels can easily cause a good image to saturate the stars. But again if it is an option you can choose and not the default that might be fine

For testing I would recommend using stellar solver tester until we are ready to test it in KStars for real

Yes, I'll follow this route. In fact, my next step on the first clear night will be to get a complete set of OSC images around the focus point, so we can experiment with them

Regards

Hi Rob,

I've been able to compile and run the new version on a Raspberry PI, but after source extraction or plate solving, I can't find any reading of the HFR to FHWD. I've assumed that was Stellarsolve who compute those magnitudes. I know that I must be missing something

Regards

So, KStars won't have the ability to use the new features yet, except that it will use the default in StellarSolver, which is now to use the Green Channel in RGB images. Also, note that these changes are in a branch, did you build from the branch? If you are in fact using the StellarSolverTester as I suggested, then you need to select that it should do the HFR in the drop down menu on the left side before doing star extraction. When you do the star extraction then, you will see the HFR next to the stars in the starlist.

Yes, don’t worry, I understand the process. I’m currently using your test program. I’ve only checked out your colourChannel branch and built the test program. I was playing with it. The detail I was missing was the HFR option in the combo. Thanks for this clue

After installation, Kstars cannot be started.

Error message:

I tried ...
Problem not solved

OS: Linux Mint 20.3 x86_64
kstars-bleeding 6:3.5.8

Any idea please?

Thanks

Tomas

Hi Rob,

I've been playing with some images and Stellarsolve test program and there is a thing that I can't understand. I've used the attached image. As you can see red/blue channels are complete different (red focused, blue out of focus), but when I use SEP w/HFR or External SExtractor w/HFR I get the same value for the HFR for each star on both channels. I expected very different results when I select red or blue channel and this is not the case. What I'm doing wrong? Because if Stellarsolve can't see any difference on HFR, the focuser algorithm will be unable to use those different channels to improve the focus point.

So the image you attached is not an RGB fits image. It is monochromatic. There are two data axes. Did you attach the wrong image or maybe merge the channels into one channel before attaching it here? Or maybe it is a bayered image and it has not been debayered?

I suspect it is a bayered image because I just tried debayering it using pixinsight and then opening it in StellarSolver and it looks ok and works as expected.

Tomas, are you using the INDI PPA? if so, then that should be automatically updated, have the latest kstars, INDI, and stellarsolver, unless there is some sort of issue? Are you using nightly or stable?

Hi Rob,

just compiling kstars against the new colorChannels version
I wonder if there's a config file for the library that selects the default handling. Or is that fixed at compile time?