astap program / Forum / General discussions ASTAP: How to match HFD octagon with batch process ?

Gert - 2022-03-21

Hi All,

I wish to do tilt analysis on a list of images. The batch processing tab 'Inspector' gives a list of hfd values (i.e. hfd11, hfd21, hfd31 ,..) How are those values correlated to the values shown in the image in the main UI when 'image inspection' is done in there? I can't find the same numbers between the two. Also would like to know the row column indexing format. (is hdf 21 top center? I.e row 1 column 2?)

I attach screen copy for my question.

Thanks & Clear Skies,
Gert

ic2177_astap_octagon.jpg

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Gert - 2022-03-21

Hi,

Appending to my question.

Is the attached row / column orientation correct?

Thanks & Clear Skies,
Gert

ic2177_astap_octagon2.jpg

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-21

Hi,

The first number is the X, the second is the Y. But for FITS normally Y goes from bottom to top. So you could reverse Y. Sou for your table, I would reverse Y.

There is a sketch in the manual but you have to read A, B, C as 1, 2, 3. See link below

The clearest indication are given by the icons for each column. The active area is darker then the rest of the 3x3 areas. But this is only corrrect if in the viewer the image flip is off both in vertical and horizontal. So it is always following the FITS pixel position. so left bottom pixel is 1,1.

Hope this helps, Han

http://www.hnsky.org/astap_hfd_areas.png

Last edit: han.k 2022-03-21

Untitled.png

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Gert - 2022-03-21

Thank you, Han for the clarification on the matrix format.

Can you clarify about the numbers between viewer and analysis table ?
(See first screen capture)

Thanks,
Gert

Last edit: Gert 2022-03-21

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-21

The both cover 3x3=9 areas.

At fits pixel position 1,1 (normally left bottom) the first area11 is located. The rest is the same.

There could be small difference between the viewer and table. The table uses a wider detection algoritm up to around HFD20 resulting in small differences.

Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-21
  
  Hello Han,
  
  This is the numbers from viewer and table. The gap is wider for larger HFD. Is there a setting to make the reports match the viewer? (what is meaning of HFD20)
  
  Viewer:
  3.60 3.77 4.34
  3.50 3.94 4.42
  3.62 3.70 4.49
  
  Table:
  3.540 4.160 4.676
  3.454 4.478 4.880
  3.549 4.092 4.849
  
  Thanks,
  Gert
  
  Last edit: Gert 2022-03-21
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-21

That is weird. The differences are too large. Can you attach or send me the image?

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-21
  
  Hello Han,
  
  Thank you for helping with the analysis. The image is here:
  https://www.skywatcher.space/download/ic2177_dat_0306_1x1_300.00_0010_c_cc_d_r.xisf
  
  In the viewer analysis I see a small gap between image in RGB and mono. But the gap to table is larger.
  
  Viewer (load xisf -> auto correct color -> image inspection)
  3.58 3.69 4.28
  3.51 3.84 4.32
  3.63 3.62 4.42
  
  Viewer (load xisf -> auto correct color -> convert mono -> image inspection)
  3.60 3.77 4.34
  3.50 3.94 4.42
  3.62 3.70 4.49
  
  Table:
  3.540 4.160 4.676
  3.454 4.478 4.880
  3.549 4.092 4.849
  
  Thanks,
  Gert
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-21

For the viewer to compare, it is important to uncheck extra stars in the menu F5

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-21
  
  Hello Han,
  
  When I uncheck the 'Extra stars' the numbers stay the same.
  (I tried both color & 'convert mono' ) (Below screen copy is mono version)
  My version is 2022.03.16 64bit.
  
  Last edit: Gert 2022-03-21
  
  ic2177_astap_octagon3.jpg
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-21

The F5 menu, so the F4 function of the viewer is smarter. In the log you can see it converts the image to mono:

Viewer (load xisf -> auto correct color -> image inspection)
3.58 3.69 4.28
3.51 3.84 4.32
3.63 3.62 4.42

The table is not so smart. (I think it does only red colour) It was intended for mono images
If i convert the image manually first to mono, then I get this:

3.554 3.710 4.281
3.489 3.861 4.333
3.617 3.616 4.398

So this looks much better. The viewer will show HFD up to value 10. The table up to HFD 20 explaining the remaining small difference.

Using debayered, so colour images for this is not so good idea. The debayer algorithm makes the HFD values larger. Better use the raw image. The F4 menu will make a nice mono image of a raw using a checker pattern filter. This is clearly missing in the table.

I will add tomorrow for the table the same image preparation of raw OSC images (and colour but better not use that) . Then the values should be very similar.

Can you send me also the raw image?

Han

Last edit: han.k 2022-03-22

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-21
  
  Hello Han,
  
  Thank you, this is great help.
  
  When you say raw image, is it OK to apply dark & flat?
  I have uploaded the dark & flat calibrated image here (but no debayer yet):
  https://www.skywatcher.space/download/ic2177_dat_0306_1x1_300.00_0010_c.xisf
  
  Thank you for the excellent work!
  
  Thanks,
  Gert
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-22

Try ASTAP version 2022-3-22. Raw and colour images should produce very similar result in viewer and tab inspector.

Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-22
  
  Hello Han,
  
  I tried the 3-22. For analysis in the viewer I balance color and convert to mono. During the second step the tool gives this error pop-up. Is there additional debug to identify the problem?
  
  Thanks,
  Gert
  
  Update : Just loading the image into viewer and convert to mono gives the error.
  
  Last edit: Gert 2022-03-22
  
  ic2177_astap_octagon4.jpg
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Gert - 2022-03-22

Hello,

There is an additional observation.
After image load into viewer I have to adjust color to make detail visible.
Then I run image inspection and the image turns green again.
At this stage when I do another color adjust the analysis graph is remove so I can't see it at the same time as the image details. Maybe I am missing a step?

Best,
Gert

ic2177_astap_octagon5.jpg

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-22

For the colours you better keep the sliders on auto, so low or medium. See attached.

For the other error I will have a look now.

Han

Last edit: han.k 2022-03-22

Untitled.png

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-22

The popup error, I can not reproduce here. Easiest way if you send me the three images and tell me what you did before the error. The debugger will tell me where the problem is.

Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-22

I found the popup error, It was is in routine converting to mono. It is now fixed, you can download again. v2022-03-22a

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Gert - 2022-03-22
  
  Hello Han,
  
  It is really quite close now.
  
  To get best match between viewer and table what should be my steps ?
  Just now I do these steps. (are all mandatory?)
  Load
  Auto correct colors
  Convert to mono (degrades match a bit)
  Adjust histogram for good data representation
  Solve
  Image inspection
  
  The best match is there when I leave the image in color in the viewer, but then it turns green after Image inspection.
  
  Thanks,
  Gert
  
  ic2177_astap_octagon6.jpg
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2022-03-22

That should not make a difference. The viewer values and the tab inspector values should be identical except that tab uses a larger search window to find stars. It has to accept larger out of focus stars. That creates a small random difference. If I make them the same then you get the same values but I can't do that because one (the viewer) is for tilt at focus and the other is for measuring tilt and curvature using in and out of focus images. You just have to except they behave a little different.

Also the stars HFD values fluctuate. That is caused by the noise. Even in simulation there is a spread in the HFD values. But having enough stars collected you get a good median value.

Cheers, Han

Last edit: han.k 2022-03-22

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

How to match HFD octagon with batch process ?

Stacking program, astrometric (plate) solver, fits file viewer

Forums

Help

How to match HFD octagon with batch process ?

How to match HFD octagon with batch process ?

Stacking program, astrometric (plate) solver, fits file viewer

Forums

Help

How to match HFD octagon with batch process ? document.SUBSCRIPTION_OPTIONS = { "thing": "topic", "subscribed": false, "url": "subscribe", "icon": { "css": "fa fa-envelope-o" } };

How to match HFD octagon with batch process ?