Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

PlanetarySystemStacker - Program features and user experience

astrophotography imaging solar planet moon
  • Please log in to reply
36 replies to this topic

#26 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 02 March 2021 - 09:10 AM

Hi Daniele,

 

As I wrote before, very small AP boxes don't make any sense. Not only does the shift detection algorithm break down, but even if that were no problem, the patch blending algorithm does not work properly. There are just too few pixels in the blending ramp, so artifacts are no surprise. That said, I did not check if your code modifications are correct. That would be another potential source for those artifacts.

 

All the best,

 Rolf



#27 airscottdenning

airscottdenning

    Viking 1

  • *****
  • Posts: 600
  • Joined: 22 Aug 2008
  • Loc: Colorado

Posted 02 March 2021 - 04:38 PM

Hi Rolf,

 

Does it make sense to autoplace large AP boxes and then "fill in" with more in regions of interest (complex topography) so we have a lot of overlap there? Or is that just a waste of arithmetic?

 

Thanks for your advice.

Scott 

Hi Daniele,

 

As I wrote before, very small AP boxes don't make any sense. Not only does the shift detection algorithm break down, but even if that were no problem, the patch blending algorithm does not work properly. There are just too few pixels in the blending ramp, so artifacts are no surprise. That said, I did not check if your code modifications are correct. That would be another potential source for those artifacts.

 

All the best,

 Rolf



#28 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 03 March 2021 - 03:49 AM

Hi Scott,

 

I'm thinking about an adaptive AP grid algorithm for a long time already. Your suggestion would be one way to go, but I would fear that the large AP boxes could only deteriorate the image quality in areas of high spatial contrast. Therefore, I would prefer to first place APs with the size selected by the user in all places with enough structure, and then fill the "holes" with larger AP boxes (e.g. in lunar mare areas). This way the large AP boxes only show up in places where the smaller boxes won't work.

 

The hope is that with those larger boxes the shift detection algorithm in low-contrast areas gets enough "grip" to compute an approximate local shift. If one just leaves the holes unpopulated with AP patches, in those areas the "background image" is used which only uses the global image shifts and does not de-warp locally.

 

All the best,

 Rolf


  • airscottdenning likes this

#29 airscottdenning

airscottdenning

    Viking 1

  • *****
  • Posts: 600
  • Joined: 22 Aug 2008
  • Loc: Colorado

Posted 03 March 2021 - 11:02 AM

Thanks Rolf,

 

If compute time isn't limiting, is there a disadvantage to just plastering high-structure areas with lots of overlapping APs? 

If I want the clearest detail in the final image and I don't care how much arithmetic I waste, is there such a thing as overdoing my APs?

 

Scott

Hi Scott,

 

I'm thinking about an adaptive AP grid algorithm for a long time already. Your suggestion would be one way to go, but I would fear that the large AP boxes could only deteriorate the image quality in areas of high spatial contrast. Therefore, I would prefer to first place APs with the size selected by the user in all places with enough structure, and then fill the "holes" with larger AP boxes (e.g. in lunar mare areas). This way the large AP boxes only show up in places where the smaller boxes won't work.

 

The hope is that with those larger boxes the shift detection algorithm in low-contrast areas gets enough "grip" to compute an approximate local shift. If one just leaves the holes unpopulated with AP patches, in those areas the "background image" is used which only uses the global image shifts and does not de-warp locally.

 

All the best,

 Rolf



#30 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 03 March 2021 - 12:04 PM

Hi Scott,

 

Thanks Rolf,

 

If compute time isn't limiting, is there a disadvantage to just plastering high-structure areas with lots of overlapping APs? 

If I want the clearest detail in the final image and I don't care how much arithmetic I waste, is there such a thing as overdoing my APs?

 

Scott

No, in principle there is no disadvantage if you do that. This "overdoing" could even result in a slight improvement because the local shifts are computed at more places. In general, however, I would not expect a big difference. The "deterioration" I mentioned above has a different reason: In areas with enough local detail you should only use the smallest AP boxes which the local shift algorithm is happy with. For smaller boxes many shift computations will fail, so the de-warping does not work properly. For larger boxes the shifts are averaged over a larger area, so local shift variations are not resolved. That is why there is a "sweet spot" in the AP size. It depends on the f-ratio and the seeing conditions, so only by comparing runs with various AP size values you can find the optimal value.

 

If you then overlap those "optimal" APs with larger ones, you unneccesarily degrade the resolution of the local warping. This is why I suggest finding the optimal AP size for the high-contrast areas and only fill the holes with larger APs.

 

All the best,

 Rolf


  • airscottdenning likes this

#31 atan

atan

    Vostok 1

  • -----
  • Posts: 111
  • Joined: 11 Jul 2006
  • Loc: San Jose, CA

Posted 19 April 2021 - 07:07 PM

I just wanted to start off with thank you for making this software. I'm using a Mac and I wasn't sure how to process my moon photos until I came across your app. I'm able to successful process video files but want to try stacking fits files instead. Is that possible or am I limited to video files only? Thanks!



#32 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 20 April 2021 - 04:20 AM

Hi atan,

 

I'm glad to read that you like the PSS software.

 

Yes, you can stack still images as well as video files. You find the details in the User Guide document on page 11. Please note that all image files must have identical pixel dimensions, and that the folder must contain the image files only. In the job specification you select the folder, and not the individual image files.

 

All the best,

 Rolf


  • atan likes this

#33 Fivel

Fivel

    Vostok 1

  • *****
  • Posts: 189
  • Joined: 14 Nov 2013

Posted 28 May 2021 - 09:57 PM

Hello Rolf. Thanks again for this fine app and your great support.

This may have already been asked, but with the latest (Mac) versions, can the app align on stars and a comet to stabilize both?

 

Thanks you.

 

Fivel



#34 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 31 May 2021 - 01:58 PM

Hi Fivel,

 

PSS is made to stack short-exposure images of planetary system objects where the seeing effects are "frozen". If the individual exposures are longer than about 1/50 sec., this "lucky imaging" is useless. Then the much simpler stacking algorithms used for deep sky imaging do a much better job. There a many programs which can be used to stack images with stars and comets.

 

All the best,

 Rolf



#35 cmas

cmas

    Explorer 1

  • -----
  • Posts: 76
  • Joined: 27 Jun 2016
  • Loc: Finland

Posted 08 June 2021 - 01:55 PM

Hi Rolf,

First of all, thanks for all the work done with developing PSS. I've read all about it here at CN and today installed "suitably old" Python and started to look at PSS a bit more with actual testing. A stacking software with proper command line options is so welcomed into the astro processing, it's one of the reasons I've started to keep my eyes open to other software than just AS3.

Item #1:
Anyway, a question regarding command line options. I typically have to stack videos using many different stacking percentages as I don't have much freetime to tweak / select the best percentage for each video. Thus, I stack using several different percentages, delete source videos to save disk space and look at the results at a later time when the family and work allows. Then I select the best stacking percentage tiffs and continue from there typically with ImPPG (which currently lacks command line options but that's another story).

I can set the percentage to be stacked in command line without any issues. But if I want to stack, let's say 20, 30, 40 and 50 %, do I need to write 5 lines of commands to a "master scripting file" or is there a more efficient way of doing things outside the obvious case i.e. for loops? In AS3 one can select 4 different percentages to be stacked and it made me think that if I write e.g. 5 lines of commands (each processing the image with different frame percentages to be stacked) PSS does everything from the start while having the option to process / store more "percentages" within one command then PSS could possibly avoid recalculating the reference frame again etc. Could I have your thoughts about this? How about running several different stacking percentages via the PSS gui? I'm just looking the most efficient way to get things done in minimal cpu time.

Item #2:
Where does PSS store log files for each session? I think that they are not stored in the same directory in which my videos are and output tiff is stored (sorry to ask such a stupid question, but I did not have a lot time to play around PSS today). Can the user specify log file directory somehow?

All in all, I have to tweak the settings etc (*) but when AS3 is lacking command line options PSS could have a lot of potential for me (lazy guy, trying to automate / standardize as much of the processing as possible) smile.gif

(*) This summer I typically image full disk solar h-alpha with asi174mm without any tracking and take videos in the range of 10-15 seconds meaning that there are quite a bit of movement within one video - this means I need to adjust some of the PSS settings to allow me to use such videos without rejecting frames due to movement. I've minimized my setup time and gear weight and in some crooked way it's actually fun to use this kind of "bare minimum" type of gear smile.gif AS3 copes with my videos nicely and I expect PSS to do the same with proper parametrizing.


Edited by cmas, 09 June 2021 - 12:23 PM.


#36 Rolf

Rolf

    Viking 1

  • -----
  • topic starter
  • Posts: 521
  • Joined: 25 Apr 2016
  • Loc: Cologne, Germany

Posted 15 June 2021 - 09:00 AM

Hi cmas,

 

Thank you very much for sharing your PSS user experience with us.

 

ad #1: If you use the graphical interface, you can go back to the stack size selection and repeat from there with a different stack size. Currently there is no way to specify several stack sizes in the command line version of PSS. Of course the command line interface could be changed, but this would require a non-trivial change to the workflow control. At the moment I don't have the plan to implement that myself.

 

ad #2: Please have a look at the User Guide document, section 4.5 "Execution Protocol". There you will find how you can produce separate protocol files for each job, and where you find them.

 

All the best,

 Rolf



#37 cmas

cmas

    Explorer 1

  • -----
  • Posts: 76
  • Joined: 27 Jun 2016
  • Loc: Finland

Posted 15 June 2021 - 09:40 AM

Hey Rolf,

Thanks for chiming in. Its a pity that the stack size cannot be set for multiple values in command line without running everything from the beginning. But that being said I fully understand your explanation and reasoning.


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: astrophotography, imaging, solar, planet, moon



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics