TheSkySearchers

RC Astro just released version 2 with AI4 today.

Just when you thought it couldn’t get any better, it just made a phenomenal leap ahead!

I just ran a quick comparison of an image I’ve been working on and not only is the faint detail better resolved, but the stars are also greatly improved.
I have some issues with CA on the brighter stars and the new BXT version significantly improved/removed the CA.

Check it out!

I have just given it a go on some images withpoorcstarsxat corners and it is just incredible. Then I tried it on some data I have from my badly collimated SCT. And it is simply astonishing. The data was basically unusable as raw, made worse by the previous BX AI and is virtually pristine with AI4. I will post some comparisons later but this is simply an incredible, quantum leap forward.

Hi Jim,

Sorry, I am not favourable on these Xterminator solutions. Why cannot someone just post their code or pixel math online ? It is like someone saying: well F=Ma, please pay me 100 dollars for this formula ! I can surely find the solution to B.urXterminator from basic image processing and statistical methods from pre-existing processes, without paying for it !


Magnus

When you go to a restaurant and order a steak, you are not just paying for the chunk of meat. You are paying for the preparer to make it "hopefully" perfect and cooked just as you ordered it.
Nobody is forcing you to buy any of the X products.

Tycho65 wrote: ↑Fri Jan 05, 2024 5:48 pm I can surely find the solution to B.urXterminator from basic image processing and statistical methods from pre-existing processes, without paying for it!

Well Magnus,
You just solved how to get it for free, follow your own advice. If you have the time and the knowledge, I say go for it. And then, if you don't mind, could you share it with us.!

Blur Xterminator is a machine learning driven AI deconvolution tool. It is not a sharpening tool that can be simply duplicated with pixelmath. The only comparison in PI is the PI deconvolution tool and that's cumbersome.

There is a ground truth in all images and atmosphere, the imaging train, and the camera will all disrupt that ground truth and convolve images. Deconvolution attempts to get as close to the ground truth as possible. I've used other deconvolution techniques in PI and CCDStack, and BXT is by far the best.

I have absolutely no problem in paying software devlopers for their product as they have to put dinner on their table like the rest of us. RC could charge everybody a subscription to use it but he doesn't. It's a one time purchase and is heads and shoulders above any other product, including PI's internal Deconvolution. Even Juan, who was skeptical, has come around.