Plate Solving issue...

Questions, support and discussions
szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Plate Solving issue...

Post by szymon » 13 Sep 2019, 12:35

Hi Han,

In the most recent ASTAP build, plate solving seems to have gotten much worse for me. I'm not sure how to go about fixing it, so here's an example file:

https://www.dropbox.com/s/ykdjm68vttgr4 ... .fits?dl=1

This file solves just fine with, for example, All Sky Plate Solver:
asps.jpg
asps.jpg (74.64 KiB) Viewed 2275 times
However ASTAP cannot seem to find a solution. I've reset ASTAP to default settings to make sure that I'm not choosing any bad settings. I'm using the same g17 star database that I used to use. I've tried explicitly specifying the field height to various settings, increasing the maximum number of stars, increasing the tolerance by an order of magnitude, and nothing I try seems to work -- ASTAP just can't solve the image (and many others from my camera).

What am I doing wrong? How can I help find a solution to this? :-)

Cheers,

-simon

By the way you have a typo on "field height":
height.jpg
height.jpg (7.47 KiB) Viewed 2275 times

szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Re: Plate Solving issue...

Post by szymon » 13 Sep 2019, 12:45

Ok, interestingly if I stack the image (just this one image in the stack) with darks, darkflats and flats, it solves just fine in ASTAP.

https://www.dropbox.com/s/06njj67w7jpmp ... d.fit?dl=1

This stack was created by trying to stack two images, default ASTAP settings, the second image was not included in the stack because of 'inconsistent solution'. After stacking, the stacked image solves no problem.

Code: Select all

11:40:40  Stack method Average
11:40:40  Oversize 0
11:40:40  Analysing images.
11:40:40  Stacking (Average), HOLD ESC key to abort.
11:40:40  Analysing darks.
11:40:40  Analysing flats.
11:40:42  Loading master dark file E:\TARGETS\EAGLE\Eagle_Ha_Darks\master_dark_20x300s_at_16C.fit
11:40:43  Loading master flat file E:\TARGETS\EAGLE\Eagle_Ha_Flats_16s\master_flat_corrected_with_flat_darks_Ha_20xF_20xFD.fit
11:40:43  Applying 2x2 flat filter
11:40:44  Adding file: 1-2 "E:\TARGETS\EAGLE\Eagle_Ha\Capture_00001.fits"  to average. Using 20 darks, 20 flats, 20 flat-darks
11:40:54  Reference image, largest with best HFD is: E:\TARGETS\EAGLE\Eagle_Ha\Capture_00001.fits
11:40:57  Adding file: 2-2 "E:\TARGETS\EAGLE\Eagle_Ha\Capture_00002.fits"  to average. Using 20 darks, 20 flats, 20 flat-darks
11:41:04  Not enough tetrahedron matches <3 or inconsistent solution, skipping this image.
11:41:06  █ █ █  Saving result 2 as E:\TARGETS\EAGLE\Eagle_Ha\Capture_00002  20xFD  20xF  20xD  0x0R  0x0G  0x0B  0x0RGB  1x300L  _stacked.fit
Finished in 31 sec. The FITS header contains a detailed history.
11:41:20  Creating monochromatic x 2 binning image to solve.
11:41:22  226 stars selected and 171 tetrahedrons selected in the image. 189 database stars required for the square search field of 1.0°. Search window at 100% based on the number of tetrahedrons. Step size at 100% of image height
11:41:22  Using star database g17
11:41:22  12 of 14 tetrahedrons selected matching within 0.005 tolerance.  Solution x:=0.472768*x+ -0.674877*y+ 895.469533,  y:=-0.675597*x+ -0.473818*y+ 2878.464007
11:41:22  Solution found: 18:18 45.4	-13d44 19	 Solved in 1.6 sec. Offset was 0.269 deg.	 Used stars up to magnitude: 12.2

szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Re: Plate Solving issue...

Post by szymon » 13 Sep 2019, 12:49

I wonder if my sensor is just too dirty for ASTAP somehow? This would go some way towards explaining why ASTAP doesn't work reliably for me as a plate solving solution in NINA (I've moved to using All Sky Plate Solver there because ASTAP was only working some of the time -- when it works its much faster but it was inconsistent).

szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Re: Plate Solving issue...

Post by szymon » 13 Sep 2019, 13:22

Another delta: the original image failed to solve via astrometry.net:

http://nova.astrometry.net/user_images/3056946#original

However the calibrated image solved just fine:

http://nova.astrometry.net/user_images/ ... #annotated

So it's not just ASTAP that had problems with it. I wonder why ASPS was fine with the original image.

Can you add the ability for ASTAP to dynamically apply a master dark and master flat before solving? A configuration option with specified calibration files, and it might also be really useful for the command line -- NINA for example could specify calibration files and platesolving would improve?

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Plate Solving issue...

Post by han.k » 13 Sep 2019, 15:28

Hello Simon,

The first uncalibrated image is 1) out of focus and 2) full of hot pixels. If view it in the astap viewer and play with the sliders, there are hundered of spikes brighter then the stars.

So first fix would be to focus better.

The temperature is 17 degrees, but I assume it hasn't a peltier cooling. But the ALTAIRH183M should have low noise Sony sensor. The noise is really strong and the hot pixels are seens as stars. You can test that by the CCD inspector.

The image is Ha image, so star are much weaker then normal. Anyhow first thing to do is to focus better. That will generate much brighter stars and more.

Do you have a automatic filter wheel? If so, maybe you could program Sharpcap to switch to L or no filter for the astrometry.

Here I have no problem with H-alpha, but my camera is cooled.

For platesolving 20 seconds should be sufficient. This image has 300 seconds.

szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Re: Plate Solving issue...

Post by szymon » 13 Sep 2019, 18:37

I agree the image isn't perfect -- but the fact remains that the calibrated version platesolves without problem, and this specific uncalibrated image platesolves fine both with ASPS and with PlateSolve2, but not with ASTAP.

That said, put that image aside (it was just an example) -- this seems to be a more general problem for me. Even under "perfect" conditions (spot on focus, CLS filter binned x2, multi-second expsoure) ASTAP is not solving well for images straight from my camera, but does solve well when calibrated. My main platesolving use case is framing in NINA. Is it likely to be an issue with my camera? Lots of really small pixels, an uncooled camera so high noise, plus a dirty sensor...I can imagine that can make it hard. That said, to work as a platesolver, I needs something capable of solving with even "bad" data...

Any suggestions for things I can improve?

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Plate Solving issue...

Post by han.k » 14 Sep 2019, 00:17

Applying a dark will certainly help. I see that for my QHY5 guide camera. Very noisy but when pHD2 subtract a dark automatically it is okay.

H-alpha imaging and an uncooled camera is not ideal.

Okay mabe something to add. Could you provide me with some short exposures e.g. 20 seconds and some corresponding darks of the same exposure length?

I could add an option to subtract a dark automatically in ASTAP.

Regards, Han

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Plate Solving issue...

Post by han.k » 14 Sep 2019, 00:56

A other option is a hot pixel filter. Stars of one pixel size.....

szymon
Posts: 28
Joined: 29 Aug 2019, 18:17

Re: Plate Solving issue...

Post by szymon » 14 Sep 2019, 10:39

Of course, I will collect some sample data for you tonight!

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Plate Solving issue...

Post by han.k » 14 Sep 2019, 11:46

Version 0.9.261 is released with calibration prior to solving. Still a few of your images, master darks, flats will be useful for my collection of test images. Good testing it essential.

Post Reply