Images solve, but only partially? (maybe camera warp/distortion?)

[Thermospore]

Hello, I am trying to calibrate a camera we set up to study the aurora!

We just need one good calibrated image, then we can use that to calculate the lat/lon of each pixel of the camera, to map the images onto the Earth

At best the center and a couple corners of the image will calibrate, but never the entire image. Perhaps the camera's field of view warps at the edges?

I looked through the parameters, but didn't see anything helpful for this

Here are the input images, the output images, and the command I ran:
```
for FILES in static_images_for_fov_calibration_2/*rescale.tif
do echo $FILES
   /usr/local/astrometry/bin/solve-field --overwrite --scale-low 45 --scale-high 160 --downsample 2 ${FILES}
done
```

Perhaps we could calibrate each quadrant of the image separately, then splice the results together?

Any insights or suggestions are appreciated, thanks!

[Dustin Lang]

Hi,

It looks like the astrometry.net code that tries to estimate the distortion is going wrong... that's probably the weakest part of the code.

I just tried adding these args, and get a decent solution.

> solve-field --config ~/data/INDEXES/4100/cfg --scale-low 45 --scale-units degw --continue --downsample 2 --crpix-center --tweak-order 3 C1_20240210_035901_280_006_rescale.tif

If you just need a single solution for the distortion, you would ideally get some stars out in the edges of the field -- we're fitting a polynomial, so it's going to do terribly if it needs to extrapolate.

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/cd404555-fa96-4060-ab06-22dd99a701ecn%40googlegroups.com.

[Dustin Lang]

ps, for visualizing what the distortion map looks like, there's a script

plotSipDistortion.py -c 100 -x 2 C1_20240210_035901_280_006_rescale.wcs distortion.png

and I get the attached plot for my example above.  (The "-x 2" is exaggerating the distortion by a factor of 2).

cheers,

dustin

[Thermospore]

Thanks so much for the explanation!

It runs like a charm now; if you are curious I attached some images of the mapping results I got with `C1_20240210_033801_546_011_rescale`!

One question: what does `--config ~/data/INDEXES/4100/cfg` do? I don't seem to have that file, so I just removed that arg when I ran it

2024年3月29日金曜日 1:00:43 UTC+9 dstn...@gmail.com:

[Dustin Lang]

Great!  That "--config" option just points to an astrometry.net config file that tells it where to find index files, and a few other settings like that.

cheers,

dustin

To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/e09a48b0-2b53-4bd8-a398-fe4729511bfen%40googlegroups.com.