How to remove Airplane trails, perform Light Pollution Removal, Vignetting Correction, Crop/Uncrop....

Hi Michel,

( I have changed your post title, I hope you don't mind )

Thank you for your questions, my answers to your questions:

1) It depends on the amount of light frames that you have. If you have more than 20 frames, in 6) INTEGRATE, use outlier rejection to remove the airplane trails. Try sigma clipping. If you have more than 50 frames, usually 1 iteration with kappa 3 is sufficient. If you have less frames, try 2 iterations with kappa 2.5 to remove the trails in your integration.

2) The boxes need to placed at areas where there is no nebulosity. So that means at areas where you think that there is only sky background. The boxes can be pretty large. And it's not a problem if the boxes contain a lot of stars. See the screenshots, the first is before gradient removal, the second after (i did the tool in this case on a single sub with both vignetting and light pollution). Try to place the boxes evenly over your image, that will help.

The flexibility slider on the left shoudl be set at the highest level for which you get a good correction. In most cases, setting it at 1 gives you the best correction. 1 means that the model calculation has strong damping which causes the model to be less influenced by possibly badly placed area select boxes. The lowest setting, flexibility 16 means no damping at all.

3) for vignetting correction the same rules apply, stars in the boxes are no problem. For vignetting correction it's quite important to place the boxes evenly over your images and make more boxes if the result isn't good. (Too little boxes could give an unstable results) Again 2 screenshots on the same light frame as in 2)

.

.

In Vignetting correction you can use three models. I have now used model

• 1) circulair, fixed center + linear gradient, this model helps with single frames which contain both vignetting and a clear light pollution gradient. This model corrects both vignetting and a linear light pollution gradient.
• 2) elliptical KangWeiss without GF (geometric factor), this model is more complex and should be used when you try to remove vignetting in an integration result or if you want to make an artifical flat for light frame calibration in the last case do this modelling on a single light frame. (by selecting create artificial flat).
• 3) elliptical KangWeiss with GF, this is the most complex model and is needed if you use optics of several optical elements. Like model 2) use this if you need to correct vignetting in an integration result or if you want to create an artificial flat from 1 single light frame for ligth frame calibration.

In both Light Pollution correction and Vignetting correction, that boxes will stay green or become yellow or even red. Red means the box could be placed wrong, you can remove it an place a new box. Don't be alarmed when there are much yellow boxes, or a couplof red ones. It's just an indication for you that the red boxes could be placed wrongly so you need to double check that the area has no nebulosity in the area box.

4) crop function: 2 options for cropping/uncropping using the batch modify tool:

1) draw a rectangle in the image for your crop and click "crop ok" (again 2 screenshots, before/after)

2) use the left, right, top, bottom offsets and click offsets ok, if you fill in positive numbers you crop (removes area), if you use negative numbers, you can add area to your frame. (4 screenshots, postive offsets before/after, negative offsets before/after)

I'll answer question 5) in a next post

Please post your questions in english, so most astrophotographers can benefit 😉 from the information.

Let me know if this information helps and if it is clear 😉

Cheers,

Mabula

Great information! I am starting to like the program more and more, especially since it is easy to use for a astro beginner like me, as long as i know where to find and how o use the tools 🙂  Thank you!

Excellent Michel, further documentation, workflows and tutorials are forthcoming 😉

Posted by: Michel

Mabula,

 And it looks like my picture is almost black and white, except for the vingetting.... how to solve this? All frames are shot with an unmodified Nikon D750 with a standard Nikon lens.

I could not attach the picture since the size is too big (and resizing costs too much detail) so i will add the dropbox link at the bottom. Since this is the first post i hope it is in the correct folder. Otherwise please let me know where to put it. And do you prefer the forumquestions to be in english or dutch?

https://www.dropbox.com/sh/bagfy31d68i9tte/AAD_q-lwO9JmyARpFowZqg-Na?dl=0

Thanks in advance,

Michel

Hi Michel,

I am looking at the data that you sent with dropbox.

I notice you created a MasterBias (MB-), 2 MasterDarks (MD-) and one master flat (MF-). Since you are using flats, a vignetting correction normally isn't needed unless the flats aren't created optimally.

I also see different exposure times in the Master Flat and the Master Darks.

Masterflat = 6.0s

Masterdark 1 = 1.7s

masterdark 2 = 120 sec

So I guess your light frames were also 120 seconds?

Possibly the flat calibration isn't working like it should. Both flats and lights need to have the bias pedestal subtracted 1 time. I think it went wrong in this case.

If you could send me 5 bias, 5 flats, 5 darks each of the different exposure lenghts that you have and 5 light frames, I will try to walk you through the workflow on how to get good calibration?

Getting colors in the image can be accomplished at the right of the image viewer, enable saturation in the DDP stretch and adjust with the SAT & SAT.TH sliders.

But I think it's best on first focussing on getting accurate calibration of your light frames 😉

Regarding saving an image in APP for publication and posting on the forum, below is a full resolution JPG created from your stack, using the save button on the right, and selecting JPG at 75% with an sRGB ICC profile. The stretch was with saturation on. The image is only 3MBs large. If you make the JPG at a higher quality % it will be much bigger in MBs. But you can post images up until 15MB at the moment. (This allows the posting of pretty high quality full resolution JPGs I think).

From the stack we can clearly see that good calibration of your light frames should be our main focus at the moment. If you send me the requested frames I'll walk you through it 😉

Mabula

I have uploaded them all in the same dropbox folder. Thanks for your help. I used 1 master bias frame which i made separate of these frames.

The saving process and colors are now fixed, i enabled saturation and saw another submenu where i could adjust some colors.

thanks for your support 

I just saw that I had 1 dark frame of only 1.7sec, which was the last frame when i stopped making them at that time. So i replaced that one in the dropbox folder as well with a good one. Same for 1 light frame, the first one of my series during setup was only 90 sec. That one is removed s well, but that one wasn't in the dropbox folder.

Hi Michel,

I have downloaded the frames,

We will need to calibrate the flat frames with the masterbias. The masterbias that you uploaded is a 32bits Pixinsight masterbias with wrong image dimensions, because PI doens't use the whole sensor for calibration due to it's DCRAW dependency. We could fix this with the batch modify tool but it's a bit too difficult for now. Can you create some new bias frame on ISO 100 (since the flat frames are iso 100 as well)? Then we should be able to get good and accurate calibration.

This masterbias of PI is the explanation why the light frame calibration wasn't working like it should. If the RAW image dimensions of the frames to be calibrated and the master calibration frame aren't equal, the calibration can't work. And PI also has lost the information on which ISO was used to create the masterbias frame. So we can't be certain this masterbias is of the correct iso for the flat frame calibration.

(It's never a very good idea to mix master calibration frames between different applications due to different implementations of Raw interpretation of the data and the calibration algorithms)

Cheers,

Mabula

Ok, thanks for the explanation. I still had some bias frames from a week later, these are now in the dropbox folder. I only put 5 in, but if you need more i have some more. These are ISO 100, 1/4000 sec. Different focal lenght, but that should not be a problem for bias correct?

thanks for all your help and information 

Posted by: Michel

Ok, thanks for the explanation. I still had some bias frames from a week later, these are now in the dropbox folder. I only put 5 in, but if you need more i have some more. These are ISO 100, 1/4000 sec. Different focal lenght, but that should not be a problem for bias correct?

 

thanks for all your help and information 

Excellent, I am going to write down a calibration workflow using your data set okay?

The focallenght has no influence on the bias frames, iso100 1/4000-th exposure is perfect.

Hi Michel,

Calibration is working now, I'll write it up in a Worflow topic 😉

One single calibrated light with masterdark, masterflat from bias calibrated flats, and a Bad Pixel Map from the darks and flats.

Mabula

nice! thank you!

Hello Mabula,

1)I tried Vignetting Correction for testing and i always get a master flat with 'a lot of circular white' to compensate the vignetting. How do i integrate this master Flat. Do i need to re-calibrate...so App will recognise the articifiel flat as a new Master Flat. What happens with the 'old' master Flat?

Does the articiciel master flat take spots (dust) away 2? I suppose not...it just corrects the pollution and optical artefacts?

Cheers

Guido

Hi Guido,

Can you open a new post and upload an image or two so I can have a better look at the artificial flat result?

The artifical flat is made based on a Kang-Weiss Vignetting model, this is a mathematical model that describes how vignetting occurs due to optical configurations. It can't remove dust spots though. For that, you need real flats. So yes 😉 it just corrects the optical artefacts.

Mabula

Hi Mabula,

In the tutoral for VC you're using one light with great result. Do we need to do this with all the lights, or can we do the same with the final stack?

Guido

Hi Guido, you can perform this in both ways:

• on all lights if you make an artificial master flat, select "create artificial flat". This will save the correction model as an artifical Master Flat. You can then use this Master Flat, just like an original master flat to correct the vignetting. To get the most accurate result, first calibrate the lights with a master dark or master bias. Save these calibrated files. Then make the artificial flat using one of the calibrated frames, and then calibrate these calibrated frames again, now with this artificial master flat.

• on the integration, but then you need to make sure you use either one of the elliptical models.

Let me know if this is clear 😉

Mabula