Can't solve this simple image

[Ferran Casarramona]

Hi,

I submitted this image that don't solve:

https://nova.astrometry.net/status/9865491

The same field of DSS colored solves well:

https://nova.astrometry.net/status/9865516

It didn't seem a difficult image.

Any idea why it fails?

Thanks,

  Ferran

[Dustin Lang]

Hmm, I don't know, with 19 detected stars, it should be fairly easy to solve.  By eye, I don't recognize how your image aligns with the DSS image -- is it rotated?

I also see that the image you submitted has some strange histogram stretch or something applied to it.  Can you submit the original?  Weird stretches usually make it difficult to correctly detect stars in the image (it's hard to estimate the noise level and rank stars by brightness).

What instrument / telescope / camera did your image come from?  Are the pixels square?  Are the 19 detected sources in your image actually stars, or are there some cosmic rays also?

cheers,

dustin

--
You received this message because you are subscribed to the Google Groups "astrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to astrometry+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/326a4e93-e84d-4059-9378-6e6cbb6c7891n%40googlegroups.com.

[Dustin Lang]

Hi,

Okay, thanks for sharing the original.

My guess would be that in all the image conversion process, your pixels end up not square.  Astrometry.net assumes square pixels, so if you picked up some horizontal or vertical stretch, then we wouldn't be able to recognize the image any more.

cheers,

dustin

On Thu, May 16, 2024 at 8:30 AM Ferran Casarramona <ferran.ca...@gmail.com> wrote:

Thanks, Dustin.

The original image is this:

https://nova.astrometry.net/status/9865408

It detects many noise pixels as stars, so tried to stretch the histogram in order to remove it.

The equipment was a 200mm newton telescope, F/4, and the camera was a Watec 910HX video camera. The image was originally 8 bits, and converted from a mp4 video format to Fits.

Regards,

  Ferran

[Dustin Lang]

Yep, that's the problem.  I just tried stretching your image, and got this result.

https://nova.astrometry.net/status/9866252

I just did

fitstopnm /tmp/1.fits | pamscale -xscale=1.036 | pnmtofits > /tmp/y.fits

on your histogram-eq image.

cheers,

dustin

On Thu, May 16, 2024 at 9:53 AM Ferran Casarramona <ferran.ca...@gmail.com> wrote:

Thanks Dustin,

Camera pixel size is 8.3µm x 8.6µm, so it's very much likely.

Ferran

[Ferran Casarramona]

Thanks, Dustin.

The original image is this:

https://nova.astrometry.net/status/9865408

It detects many noise pixels as stars, so tried to stretch the histogram in order to remove it.

The equipment was a 200mm newton telescope, F/4, and the camera was a Watec 910HX video camera. The image was originally 8 bits, and converted from a mp4 video format to Fits.

Regards,

  Ferran

El jue, 16 may 2024 a las 14:17, Dustin Lang (<dstn...@gmail.com>) escribió:

[Ferran Casarramona]

Thanks Dustin,

Camera pixel size is 8.3µm x 8.6µm, so it's very much likely.

Ferran

[Murray Forbes]

I also have a video camera (WATEC 120N+) that has 8.3 x 8.6 pixels and couldn't solve anything until the image was stretched in the horizontal axis to make the pixels square.

There's a conversation about this in this group (Plate solving fails for Watec 120N+ video camera images).

Murray.