×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

New KStars/Ekos Module: Analyze

  • Posts: 1208
  • Thank you received: 559
Thanks Scott!

I think all those are already covered.
invent.kde.org/education/kstars/-/blob/m...yze/analyze.cpp#L151
However, if you notice your filters not being color-coded in Analyze, please let me know.

Hy
3 years 6 months ago #60507

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
@schwim

Greg,

I guess you're right that sky background normalizes by the max value.

Quick HFR shouldn't make a difference in seeing the values (let me know if it does). What it does is just compute HFR on the center 25% of the image. That speeds up the computation considerably, as in some cases the hfr can take quite a while to compute on a raspberry pi. If you have auto compute hfr on, but yet hfr is not being computed, please send me a log with debugging turned on for at least the EKOS sections: Indi, FITS and Capture, and Verbose & File checked at the top of the logging dialog.

BTW, the value might read 0 if you have the cursor set to somewhere outside of a capture interval. Click on a capture in the timeline (one that's after you turned on auto-compute HFR, of course) and check the value displayed as well the value printed in the "Details" box in the lower left, and let me know if it's still 0.

Thanks,
Hy
3 years 6 months ago #60509

Please Log in or Create an account to join the conversation.

  • Posts: 152
  • Thank you received: 20
Hy,

That would be a fail on my part if that were the case! Alas, it's not cursor placement. If I go outside of a sub it will show blank in many cases but not all. If I click a sub, it shows 0.00. I opened the .analyze file and see the HFR as 0.00 in all CaptureComplete lines, and if I manually put a number in there it does reflect in the graph, so the graphing appears to be working correctly.

I'll do some more debugging this evening and let you know what I find.

Thanks!
Greg
Last edit: 3 years 6 months ago by Greg.
3 years 6 months ago #60511

Please Log in or Create an account to join the conversation.

  • Posts: 152
  • Thank you received: 20
Hy,

OK, whatever I poked at yesterday now provides an HFR number in the Analyze tab. However, the number never changes from sub to sub, and the HFR calculated in the fits viewer is significantly different than the number presented on the graph (e.g. 1.34 in viewer, 0.94 in Analyze). Focus HFR is also different. I tried changing the Quick HFR viewer setting but that did not change things.

Greg
3 years 6 months ago #60559

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
There's a log line (if you have fits verbose logging set) that looks like this
[2020-04-27T03:21:10.600 PDT DEBG ][           org.kde.kstars.fits] - FITS HFR: 2.35749
It's printed to your kstars log file when the HFR is computed.

There's also, of course, the HFR in the .analyze file in 'CaptureComplete' right after the filter and before the filename:
CaptureComplete,9043.158,240.000,Green,1.499,/home/pi/mydisk//seahorse/Light/Green/seahorse_Light_Green_240_secs_2020-09-19T22-28-33_223.fits
Those should have the same value for the same capture (the above examples are not from the same capture ;)

Are you seeing differences? Is it plotting something different?
If you want me to have a look, please send the log and the .analyze.

Thanks,
Hy
3 years 6 months ago #60572

Please Log in or Create an account to join the conversation.

  • Posts: 152
  • Thank you received: 20
I don't have fits debugging turned on. Will give it a shot again tonight. At the very least the focus HFRs are not matching. The analyze data looks like this:
CaptureComplete,6910.911,300.000,L,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/L/Light_L_300_secs_2020-09-23T23-06-30_001.fits
CaptureComplete,7830.227,600.000,L,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/L/Light_L_600_secs_2020-09-23T23-20-35_001.fits
CaptureComplete,8458.568,600.000,L,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/L/Light_L_600_secs_2020-09-23T23-35-01_002.fits
CaptureComplete,9530.181,900.000,L,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/L/Light_L_900_secs_2020-09-23T23-52-53_001.fits
CaptureComplete,11028.627,1200.000,L,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/L/Light_L_1200_secs_2020-09-24T00-15-44_001.fits
CaptureComplete,13210.166,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T00-54-13_015.fits
CaptureComplete,15094.138,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T01-25-37_016.fits
CaptureComplete,17134.684,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T01-59-37_017.fits
CaptureComplete,19174.548,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T02-33-37_018.fits
CaptureComplete,21212.947,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T03-07-35_019.fits
CaptureComplete,23264.680,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T03-41-47_020.fits
CaptureComplete,25316.347,1800.000,Ha,0.940,/home/schwim/Dropbox (7e7)/Astrophotography/working/fsq106/ngc7000/Light/Ha/Light_Ha_1800_secs_2020-09-24T04-15-59_021.fits
3 years 6 months ago #60574

Please Log in or Create an account to join the conversation.

  • Posts: 163
  • Thank you received: 26
Hello Hy, great idea and excellent work for this module!!

I was wondering if you could 'just' use text for the filters, instead of the close-to-each-other colours? :)
And maybe a toggle switch for the guide graph to switch between 'time / frequency' (using Fast Fourier Transform) so it's relatively easy to analize if there's an obvious periodic error in the mount.
Last edit: 3 years 6 months ago by Bart. Reason: spelling.
3 years 6 months ago #60575

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Obviously, it looks suspicious that all the HFR values are 0.940.
It would be good to know if the system is computing them wrong, or if they are being recorded wrong.
Did you say that they were being displayed correctly on the fitsviewer?
It will be hard for me to debug -- so the more info you can provide (and of course the more debugging you can do, if you're comfortable with that)
the better.

BTW, what kind of camera, etc?
Hy
3 years 6 months ago #60576

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
@bart

I considered the FFT idea, certainly could be a future. If you want to see that now, you may not know that you can take the guidelog that is saved (see the guidelogs directory that's parallel to the logs directory). That log is compatible with phdlogview, even if you were using the internal guider. So load a log into phdlogview, and use their frequency analysis tool.

Hy
3 years 6 months ago #60577

Please Log in or Create an account to join the conversation.

  • Posts: 163
  • Thank you received: 26
I don't need it. I use Python for data analysis so definately am comfortable with fft, but maybe as an idea for a future indeed. And if I can help you, I'll be happy to!
3 years 6 months ago #60579

Please Log in or Create an account to join the conversation.

  • Posts: 152
  • Thank you received: 20
The fits viewer does give numbers that are a bit different and seem more correct. I'm happy to debug.

The camera in this case is an FLI ML16000.
3 years 6 months ago #60580

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Replying to @dokeeffe from his comments on indilib.org/forum/general/7863-the-new-analyze-tab-thanks.html

Derek,

Thanks again for trying out Analyze. I agree that an SQR-type measure is something that should be in Analyze.

SQM itself is possible, but there are complications . For example, I believe the right way to do it is to subtract a dark from the signal,
get the median pixel value, multiply by the inverse of the bandwidth of the optical system, (e.g. if you're using a Blue fliter, or more extreme Ha,
fewer photons reach the sensor than if you're using an L, or a 1-shot camera) then do a little math--have it written down somewhere ;)
This is complicated by the need of the dark and the need to know the filter bandwidth, but certainly possible.

I did provide "Sky Background", the checkbox furthest to the right on the 2nd line, which is the background sky value SEP (Sextractor)
computes when extracting stars, but I see yours isn't populated. I think you need to be using SEP MultiStar guiding to get that value.
I can improve requirements like this after 3.5 when SEP improvements being done by @rlancaste are scheduled for release
You might SEP MultiStar guiding and look at the sky background feature if possible.

Another straight-forward thing I could do is simple add the median pixel value, which is "pretty much the same" as the SQM, except of course
it's linear, not log, has an offset, is affected by filter bandwidth, ... ;) I have gone through the exercise of computing SQM from median pixel values.
Obviously SQM would be more ideal, though.

Adding new features (now that I've filled up the 2 rows) would required a little UI redesign, but I don't think it's too bad. I had planned on organizing
these features into "guiding", "image" and "mount" sections, and SQM and ellipticity would fit in the image section nicely.

What do you think? Would the basic median work for you? Does sky-background look OK as is?

Hy

PS I see you're afflicted with the same "after focus guiding correction" that afflicts me.
I'll definitely look into this further someday, if someone doesn't beat me to it.
The following user(s) said Thank You: Derek
3 years 6 months ago #60736

Please Log in or Create an account to join the conversation.

Time to create page: 0.521 seconds