Star Alignment

[Emanuele]

Hi all,

I am new to the Pixinsight platform. I have been using the common tools (Maxim, PSCS3 etc).

I am playing with a set of data, of the veil nebula - very wide field - and trying to star align the frames. They have been calibrated.
StarAlign tells me it succeeded in aligning the frames, but the result shows to opposite. There's a very minuscule mismatch in the alignment and stars have tiny trails - it even looks like field rotation.

There must be something I am doing wrong?
Thanks!
Emanuele

[Emanuele]

...oh , I was forgetting: I also cannot seem to be able to remove hot pixel from the image? Is there an outliers detection in PI that can automatically remove hot/dead pixels like maxim?

[Emanuele]

...aaaaaand back to the usual programming. I had skipped the Sigma Clipping step and the poor PI was trying to reject nothing      I need a vacation.
Still I am not sure why Star Alignment is not working as it should. It works in other images, but not with this very wide field.

Any ideas why?

[Juan Conejero]

Hi Emanuele,

Welcome to PixInsight Forum. Don't worry, all of us --including me very especially-- need a good vacation for sure.

There's a very minuscule mismatch in the alignment and stars have tiny trail

That should never happen. Can I see an example of this problem?

Could it be that the images have a large amount of hot pixels, which are not being removed by StarAlignment's automatic hot pixel removal? If this happens, then SA could be aligning with respect to the hot pixels, instead of the stars. Try increasing the value of the hot pixel removal parameter (Star Detection section). The default value is 1 pixel; try with 2 pixels.

Is there an outliers detection in PI that can automatically remove hot/dead pixels

Indeed. You'll love our ImageIntegration tool when you give it a try. To get your feet wet, take a look at this video tutorial by Harry Page. The new version (to be released really soon) comes with even more pixel rejection algorithms and new features.

Besides ImageIntegration, you also have Nikolay Volkov's cosmetic correction script and Carlos Milovic's DefectMap module (the latter will be integrated in the ImageCalibration module in the next version).

[Juan Conejero]

It works in other images, but not with this very wide field

Have all of the images been acquired with the same lens or telescope?

Keep in mind that StarAlignment cannot correct for dissimilar field distortions when they are very large, as usually happens with wide field images. For example, if you try to align one image taken with a 50 mm lens and another with a 28 mm lens, you'll probably have problems because the field distortions caused by both lenses are very different. SA will register the central areas of both images very well but it probably will ignore all stars on the peripheral areas. This happens because SA uses triangle similarity as its primary star pair matching method, and dissimilar field distortions don't preserve it, in general. For these difficult cases, DynamicAlignment is the tool of choice in PixInsight. DA is a semi-manual image registration tool able to align virtually any pair of images, despite their differences in FOV and field distortion.

I would need to take a look at one pair of those problematic images to say something more specific.

[Emanuele]

Juan,

thank you for your help!

The images have all been taken with the same telescope and camera. I am trying a new alignment with the suggestion you have made.
I'll post the results soon.

[Emanuele]

Alright then.... there's no need to post the results. Your advice to increase the hot pixel removal parameter to 2 fixed the issue I was having. Thank you!

One more question which is not related to this at all: I accidentally move a process window (Image Integration) too high and the top bar is behind the OSX menu bar. Because of this, I cannot move the process window anymore! Is there a way to restore the process windows to a particular position on the screen?

Thanks!
E.
 

[Niall Saunders]

Emanuele,

I have seen my neighbour suffer from this - I think that the solution that you are 'missing' is that - when this occurs - your PI main window will have acquired a Vertical ScrollBar at the right-hand edge (Sometimes you can also acquire a Horizontal ScrollBar as well).

Use the Vertical ScrollBar to 'find' your missing 'top bar', and all should be well again.

I hope this is what you are suffering from

Cheers,

[mmirot]

Indeed. You'll love our ImageIntegration tool when you give it a try. To get your feet wet, take a look at this video tutorial by Harry Page. The new version (to be released really soon) comes with even more pixel rejection algorithms and new features.

Besides ImageIntegration, you also have Nikolay Volkov's cosmetic correction script and Carlos Milovic's DefectMap module (the latter will be integrated in the ImageCalibration module in the next version).

I hope you will included this script and module in the next release!

Max

[Harry page]

Hi

Welcome to the mad house

You have not been taking your video pills have you 

But now you know what to do I bet you are pleased with the results 

When mr Juan releases this new module , I will have to remake my video , will the work never end 


Harry

[Emanuele]

Emanuele,

I have seen my neighbour suffer from this - I think that the solution that you are 'missing' is that - when this occurs - your PI main window will have acquired a Vertical ScrollBar at the right-hand edge (Sometimes you can also acquire a Horizontal ScrollBar as well).

Use the Vertical ScrollBar to 'find' your missing 'top bar', and all should be well again.

I hope this is what you are suffering from

Cheers,

Hi Neil!
The attached screenshot show what the problem is! I'm stuck! 

Harry: thanks for the welcome! I love it here Lots to learn. I hope to have some data to play with, soon. Unfortunately the weather here in Italy has been terrible since last November. To the point where I am considering shipping everything to a remote - dark - location.

[Niall Saunders]

Hi again Emanuele,

Sorry, I can't tell much from that 'cropped' screenshot - in other words, I can't tell whether the 'vertical scrollbar' has appeared at the side of your main PI screen or not.

I also can't 'simulate' the behaviour that I 'think' you are having, because I am now running a triple-monitor setup, and I can drag my Process Dialogue boxes 'out' of the PI 'main' window, and drop them elsewhere on my triple-monitor desktop. And I seem to remember having had to 'configure something' to be able to achieve this.

However, I even wonder whether you are also working in a similar mode. Can you try setting your PI 'main window' such that it is NOT maximised on your desktop, and then see what happens when you push a dialogue window 'off the edges'. Does it actually 'move out', onto the desktop itself?

Cheers,

[Juan Conejero]

Hi Emanuele,

You can click and drag on the bottom bar of a process window to move it.

Anyway, sorry for the inconvenience. The next version fixes these problems completely on all platforms; windows can no longer be placed outside the visible area of the desktop in that way.

[Emanuele]

Thanks Neil and Juan.

I finally fixed it It was my mistake, just careless moving around of processes windows trying to understand PI! lol