Planetary

Farzad_k

Well-known member
Hello,

Just wondering if features of PI can be used to align and stack objects like the moon or other planets. I am keeping my fingers crossed hoping someone will say yes.
 
there is a script called FFTRegistration that works reasonably well on the moon and the sun.

i think people have even managed to make StarAlignment register images of the moon, essentially tricking it into 'thinking' bright features on the moon are stars. there's a thread here somewhere about that.

rob
 
Perhaps future development of PixInsight will enhance its capabilities in this area - especially given that the much-loved "Registax" package is no longer being supported (or so it would seem) by Cor Berrevoets.

Whilst I would never want to take anything away from Cor's tremendous efforts over the years, much of the pre-alignment and post-alignment processes offered in Registax are already available (to some extent) within PixInsight - and the actual FFT-based registration process is not totally outwith the capabilities of PI either.

So, come on all you coders out there - who fancies rolling up their sleeves and putting together a PixInsight PJSR script that can duplicate the stages implemented in Registax?
 
pfile said:
there is a script called FFTRegistration that works reasonably well on the moon and the sun.

i think people have even managed to make StarAlignment register images of the moon, essentially tricking it into 'thinking' bright features on the moon are stars. there's a thread here somewhere about that.

rob

Thanks. I will give it a try.
 
Niall Saunders said:
Perhaps future development of PixInsight will enhance its capabilities in this area - especially given that the much-loved "Registax" package is no longer being supported (or so it would seem) by Cor Berrevoets.

Whilst I would never want to take anything away from Cor's tremendous efforts over the years, much of the pre-alignment and post-alignment processes offered in Registax are already available (to some extent) within PixInsight - and the actual FFT-based registration process is not totally outwith the capabilities of PI either.

So, come on all you coders out there - who fancies rolling up their sleeves and putting together a PixInsight PJSR script that can duplicate the stages implemented in Registax?

It would be really nice if PI could develop a batch process or a script that was exclusive to lunar image processing. It seems all the software that are currently out there, free software, are either abandoned or unsupported and pretty much all of them lack a useful set of documentation.

Farzad
 
Hi
I use Star Alignment for aligning my lunar images acquired with my DSLR; and it works most of the time. Under Star Detection chose a high value for peak response, f. ex. 0.90 and higher; and a low value under maximum distortion, f.ex. 0.10.
For aligning and stacking images acquired with my planetary camera, I use a different program.
Hope this can help you.
Raymond
 
I'm glad I ran across this post.  I had trouble stacking my DSLR lunar images with Siril or AutoStakkert.  FFTRegistration did a very nice job!
 
Hi,

I'm glad to know the old FFTRegistration script is still useful. I wrote this script back in 2005. Wow, that's 12 years ago!

A new tool for feature-based image registration (probably using SURF descriptors, or maybe our own descriptors) is on the drawing board. It will include a high-performance arbitrary distortion correction algorithm similar to the next version of the StarAlignment tool, which will be released after the initial PixInsight 1.8.5 version. This new tool will be a primary building block for panorama generators in PixInsight. It will be also great for registration of lunar images and generation of lunar mosaics.
 
That is great to know. We appreciate PI being able to process planetary imaging especially lunar.

Registration (alignment) being part of the whole process, it would be great if someone could come up with a total process for planetary imaging because in some cases the image is dancing back and forth due to atmospheric conditions.
 
Hi,

The current gold standard for stacking planetary/solar/lunar images is AutoStakkert. There really is no reason to use a swiss army knife to drive in a nail when a perfectly good hammer is available for free. After that you will of course process your stacked image in PI.

The above is for Windows users. Not sure what the state of the art is for Mac/Linux.
 
I disagree that there are any useful standards out there let alone a "god standard". People do talk about the software(s) that are available but the free software that are available are pretty much abandoned (Registax) or are poorly documented (PIPP and Autostakkart) and suffer from stability issues (Registax, PIPP, Autostackkart). Yes, some people have eventually figured out when to use which one of these tools in a rhythm that works for them, but that doesn't make any of the available software "god standard" unless by that you mean the only tool available to use.

Swiss army knife or a hammer - something that is free is free for a good reason. One common issue I have noticed with these software is that they like to work on very small images, and I don't, I don't care for 400x 400 pixel images of the moon or Saturn or Jupiter. I want to be able to use my full frame camera, capture with as high a resolution as possible, and be able to print as large as possible.

So please - PI - give us a tool that is worthy of our passion and the thousands of dollars worth of equipment that we have gathered together to capture the images with.
 
"One common issue I have noticed with these software is that they like to work on very small images, and I don't, I don't care for 400x 400 pixel images of the moon or Saturn or Jupiter"

Gosh. And here I have used AutoStackert to stack full frame ASI1600 images of the moon. But please enlighten us with all your years of experience.
 
AutoStakkert works well using Wine in Linux.  I use Crossover Office, and I expect that Mac users would be able to do the same to run AutoStakkert.

I had a series of DSLR files, and that was what didn't stack well in AutoStakkert.  It stacked my SER files from my ASI 224MC quite well.
 
Nocturnal said:
"One common issue I have noticed with these software is that they like to work on very small images, and I don't, I don't care for 400x 400 pixel images of the moon or Saturn or Jupiter"

Gosh. And here I have used AutoStackert to stack full frame ASI1600 images of the moon. But please enlighten us with all your years of experience.

I never said I had years of experience. I have not been able to get Autostakkart to run or it looks like it is running but it is not doing anything. Searching on Google it turns out that is a random issue some have found to go away by re-downloading and reinstalling. I have noticed, in my little time with Registax, that when I crop my files down to 400x400 it manages to work. And PIPP has similar issues.

Here is what I want you to consider: this subject was initiated on PI as a question about PI features. It is not intended to challenge you and your kind or argue whether the crappy software that is available is really crappy or not. I have made a suggestion. You don't like it? move on.

Thank you.








 
RDBeck said:
AutoStakkert works well using Wine in Linux.  I use Crossover Office, and I expect that Mac users would be able to do the same to run AutoStakkert.

I had a series of DSLR files, and that was what didn't stack well in AutoStakkert.  It stacked my SER files from my ASI 224MC quite well.

I use PC machines and I know from my limited time being involved with these freebies that there are definitely more than one way people are finding them to be operational. Something funny someone suggested to me when I was asking for help getting Autostakkart to work was to run my video through PIPP as it somehow fixes a problem. And when you install PIPP it recommends installing a codec that you don't need but Registax does. It is very convoluted.
 
Juan Conejero said:
Hi,

I'm glad to know the old FFTRegistration script is still useful. I wrote this script back in 2005. Wow, that's 12 years ago!

A new tool for feature-based image registration (probably using SURF descriptors, or maybe our own descriptors) is on the drawing board. It will include a high-performance arbitrary distortion correction algorithm similar to the next version of the StarAlignment tool, which will be released after the initial PixInsight 1.8.5 version. This new tool will be a primary building block for panorama generators in PixInsight. It will be also great for registration of lunar images and generation of lunar mosaics.

Good to see more than deepsky processing tools on the roadmap, (almost) everyone wants to make some solar system images also  :)

The panorama generator can that help to change a fussy looking tree (due to StarAlignment) in a sharp one?
 
Found this thread as I own Pixinsight and want to use it to process planetary .avi's

Still searching for the answer but...

So funny to see people raving about other software when, as has been pointed out, this thread was about Pixinsight functionality - not AutoStakkert.

And calling that a hammer (instead of a multifunction knife) is hilarious - AutoStakkert is at best a noodle - you'll spend your life getting it to work - we want PI to provide this functionality so our time and expertise are localised in software - duh.....

David
 
FFTRegistration in PI works very well for aligning lunar images, and the processing tools PI offers allow for excellent lunar image processing.

What's missing in my opinion is the video frame analysis to select the very best frames for alignment and stacking.

Here's an example from last night: stack of 30 x 0.003 second frames at full resolution with my ASI1600 at 1050 mm. Unfortunately the forum requires that attachments be smaller than 0.5 MB, so resolution is degraded here, but honestly PI is a GREAT tool for these images. Now we just need a way to extract and select frames from a video file.
 

Attachments

  • Moon.20190216.0.003s.jpg
    Moon.20190216.0.003s.jpg
    498.9 KB · Views: 116
FWIW, I have had great luck aligning moon images using StarAlignment with the following settings:
--Detection Scales: 6
--Noise Scales: 4
--Noise Reduction: 4
--Max Distortion:0.1
--RANSAC Tolerance: 6
--RANSAC Iterations: 3000

Although I am very interested if Juan can give us any follow up to his comments from June 2017 about a "new tool for feature-based image registration". He described this tool as a "primary building block for panorama generators which would also be great for registration of lunar images and generation of lunar mosaics." Is this tool available??
 
Back
Top