It might be kind of cool, if solving fails, it falls back to another method automatically with the last resort being the online blind solve.

I am attaching an image. I need to bring up the Rpi to get another, maybe later.

Hmm, not letting me add a .fits file?

Ron

I will have to do something else?

Ron

You can upload zip files with fits images. Or change the file extension to something the forum accepts.. fits would be best if that’s what you were testing. If you change it to jpeg data will be lost from the fits headers. That info may be key to solving it or key to why it was not solving .

Yes I am already working on this for the next version.

You could upload the file(s) to nova.astrometry.net/ and give the link or use this site:

uploadfiles.io/

And give the link.

p.s. the G18 is about 1 gbyte. Not 10 gbyte

Han

Sorry, I just realized that what I said yesterday was not clear, particularly the part about FITS files and jpeg files.

I meant to say that resaving the file in a format that the forum accepts is not a desirable option since we would like to preserve the FITS headers and see what its going on. One option that would allow you to submit the FITS file(s) would be to take the fits files that would not solve, put them all into a zip file and submit that file. Another option would be to just take the file extension and change it to one that the forum would accept, and to let us know so that we can change it back after we download it, since that doesn't actually change the contents of the file. Hopefully this is more clear.

Han,
You are right, When I downloaded the indexG18 zip file I was worried it might use up a lot of memory so I ran "unzip -l" on it and I as sure it said 10G uncompressed. But now that I look at it on the USB memory stick the G18 library is only 1 G in size. So I can put it the SD card.

thanks,
Ron

Please bear with me as my experience with plate-solving is limited to running it in Ekos during polar alignment or slewing to a target.
I am running on an Rpi 4, Raspbian "Buster" Version 10; Kstars 3.40, Build: 2020-01-06T10:36:16Z (I have this working with all equipment after working out some 'bugs' in the setup, so I don't want to upgrade right now - although I do a clone backup SD card)

I have ASTAP with the G17 and G18 index files loaded. I recall using this local solver succesfully in the past, but as I indicated with a wider FOV setup.
I have Astrometry.net solver loaded (solve-field) with the set of index files that Ekos recommended for this f/10 setup.
And now I have Sextractor and SexySolver-Tester loaded on the Rpi.

This is an image from an imaging session on May 7th. I captured this image and saved it, then captured another after starting the polar alignment routine, the second image solved using online Astrometry.net in 102 seconds
"Solver RA (280.54) DEC (80.339) Orientation (275.58) Pixel Scale (0.48818)" the RA/Dec values seem strange since after completing the first round of the PA routine it said I was off the NCP by 3 1/2 degrees.

I just uploaded the first image file to here: May7th fits image

I also just now uploaded this file to Astrometry.net and it started processing the file and it failed?

I have run this image thru ASTAP with both the G17 and G18 star databases and both failed. I think ASTAP only used 100-200 stars.

I am now trying SexySolverLibrary Tester. It says it found 3525 stars (it must be picking up a lot of noise or hot pixels - I did not do any dark subtraction on this image)
I am running it using the internal Sextractor and Internal SexySolver. I have not explored all the other options but I did set the scale arcminwidth to 20 low and 30 high.
it is running a long time over 300 seconds, so I suspect it will fail as well.

I will look to find another image that I can get to solve using Astrometry.net online, like it usually does during observing and try these other solvers.

As I said I am a 'little' lost here. I gather using this tester I can try solving an image using the internal program, ASTAP, local Astrometry.net as well as the online versions of some of them.
It would help me if you can give me a set of directions of how to proceed. Otherwise I will need to read thru the 15 pages of this forum topic to see what others have done.

thanks,
Ron

Ron,

I didn't solve this file in ASTAP or with SexySolver and Astrometry either. I did, however, get a reported solve in ~1/2s using SexySolverTester 1.3 on my desktop, which is a pretty fast processor using the Internal SexySolver with the ParallelSmallScale option and the "Fast Solving" option. However the RA & DEC ERR values were large for both options:-(. Rerunning both options with "Use Position" (Scale/Position tab) solved in ~0.6s with ERR values of 0 and the same RA & DEC values as the other solver.

Richard

Hi, so I just ran your test image, it solved in about a half second if I selected the "FastSolving" Profile in the Internal SexySolver. It also solved with my Ext Sextractor, Ext Solver method. You are correct that a number of the other solving methods failed. I had it solve with a 4204 index file, and a 4205 index file.

The Indicated error in RA and Dec from the values in the file were indeed very large. I think the solution that I found was close to what you had said it should be Ron.

One issue I think you might have here is that not only is your field very small, but there aren't really that many stars. You are correct that the 3,000 something stars is not accurate. Thats why I would use the FastSolving profile. It purposefully sets some parameters to filter out "stars" that aren't stars or aren't useful in solving.