Stacked colour fits

Questions, support and discussions
Serg1
Posts: 37
Joined: 30 Sep 2018, 05:14

Stacked colour fits

Post by Serg1 » 28 Dec 2018, 13:53

Hello Han.
When I summed fits from QHY8L (colour) and Convert OSC images to colour not checked, Method Bilinear interpolation, Star alignment resulting image is bad. See picture.
When Convert OSC images to colour is checked, then fits not stacked.


16:49:33 Stack method Average
16:49:33 Oversize 0
16:49:33 OSC, demosaic method Bilinear interpolation
16:49:33 Analysing images.
16:49:36 Average HFD is 3,5
16:49:36 Stacking (Average), HOLD ESC key to abort.
16:49:36 Analysing darks.
16:49:36 Analysing flats.
16:49:36 Loading master dark file F:\tempfoto\yan_may_2018\dark\master_dark_1x10s_at_-29C.fit
16:49:37 Loading master flat file F:\tempfoto\yan_may_2018\flat\master_flat_corrected_with_flat_darks__20xF_20xFD.fit
16:49:37 Adding file: 1-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_10s_20180215_031501.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:38 Reference image, largest with best HFD is: F:\tempfoto\yan_may_2018\20Cep\20Cep_10s_20180215_031501.fits
16:49:39 Adding file: 2-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_10s_20180215_031520.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:40 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:40 Loading master dark file F:\tempfoto\yan_may_2018\dark\master_dark_1x10s_at_-29C.fit
16:49:41 Adding file: 3-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_10s_20180215_031539.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:42 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:42 Adding file: 4-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_10s_20180215_031558.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:44 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:44 Loading master dark file F:\tempfoto\yan_may_2018\dark\master_dark_1x60s_at_-29C.fit
16:49:44 Adding file: 5-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_60s_20180215_031009.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:46 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:46 Adding file: 6-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_60s_20180215_031117.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:47 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:48 Adding file: 7-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_60s_20180215_031226.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:49 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:50 Adding file: 8-8 "F:\tempfoto\yan_may_2018\20Cep\20Cep_60s_20180215_031335.fits" to average. Using 1 darks, 20 flats, 20 flat-darks
16:49:51 Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
16:49:52 █ █ █ Saving result 8 as F:\tempfoto\yan_may_2018\20Cep\20Cep_60s_20180215_031335 20xFD 20xF 1xD 0x0R 0x0G 0x0B 0x0RGB 1x60L _stacked.fit
Finished in 20 sec. The FITS header contains a detailed history.
Attachments
2018-12-28_16-36-06.png
2018-12-28_16-36-06.png (404.6 KiB) Viewed 13939 times

han.k
Posts: 288
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Stacked colour fits

Post by han.k » 28 Dec 2018, 15:51

Hello Serg,

Send me or drop somewhere some images, so I can have a look? Your screen shot indicated enough stars, so it should work.

Han

Serg1
Posts: 37
Joined: 30 Sep 2018, 05:14

Re: Stacked colour fits

Post by Serg1 » 28 Dec 2018, 18:44

See private. Bayer pattern RG.

han.k
Posts: 288
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Stacked colour fits

Post by han.k » 28 Dec 2018, 20:36

Hello Serg,

I have downloaded them. RG seems the wrong de-Bayer setting. Hardly and red color left. In the program I use the red channel only to stack so that's why the program can't find stars to align. I assume that BG is the correct setting for de-Bayer.

With BG setting, strange enough the program freezes after the second image. Something goes wrong with these OSC images. So this is an interesting test set. I will start a deeper investigation and hopefully have a solution very soon.

Han

han.k
Posts: 288
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Stacked colour fits

Post by han.k » 29 Dec 2018, 00:17

I found the hanging problem. The histogram was calculated for all colours while star detection uses the red values. Since your images are very red, the star detection routine is testing each pixel using a slow routine.

I fixed it in the Windows zipped version only. I changed a lot, so more testing/changes are required:

http://www.hnsky.org/astap.zip

Tomorrow I will continue.

Regards, Han

Serg1
Posts: 37
Joined: 30 Sep 2018, 05:14

Re: Stacked colour fits

Post by Serg1 » 29 Dec 2018, 04:14

Hello Han.
RG is correctly. The image was taken in the morning. The sky was blue.

han.k
Posts: 288
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Stacked colour fits

Post by han.k » 29 Dec 2018, 08:32

Hello Serg,

Okay, I will continue testing/"code optimizing" today. Images from a different setup are always good for testing.

Han

Serg1
Posts: 37
Joined: 30 Sep 2018, 05:14

Re: Stacked colour fits

Post by Serg1 » 29 Dec 2018, 11:17

Hello Han.
I propose an algorithm:
1. de Bayer.
2. Convert RGB to grayscale.
3. Alignment.
4. Addition is channel-by-channel RGB.
5. Repeat for the next image.

han.k
Posts: 288
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Stacked colour fits

Post by han.k » 29 Dec 2018, 12:05

Hello Serg,

Fixed it in version 0.9.150. This version works with your images and de-Bayer as RG.

I corrected a bug in my OSC code. I also overhauled the histogram. Now it shows colours.

I have to do some more testing. Changed a lot in the code so be aware of bugs.

The process is now:

1. de-Bayer.
2. Alignment on red channel.
3. Repeat for the next image.

At the moment I still prefer to align on red. This makes the code simpler for mono and colour images.

I'm also a little puzzled about the colour of the brightest stars in your image.

Serg1
Posts: 37
Joined: 30 Sep 2018, 05:14

Re: Stacked colour fits

Post by Serg1 » 29 Dec 2018, 17:44

Hello Han.
Get better. But there were artifacts in the image (#1). In the super-pixel method, they are more pronounced (#2).
Attachments
2018-12-29_20-30-54.png
#1
2018-12-29_20-30-54.png (71.88 KiB) Viewed 13898 times
2018-12-29_20-37-09.png
#2
2018-12-29_20-37-09.png (83.92 KiB) Viewed 13898 times

Post Reply